# Hello, World! Tech Entrepreneur | AgTech, Computer Vision | Startup Advisor | 25+ yrs in tech

## Alex Samson, a tech entrepreneur Tech entrepreneur with 3x exits and 10x fails. Supporting [AgTech innovation](/agitech) since 2015 through R\&D and product development. [Mentoring](/startup-mentor) early-stage founders in 500 Global and other startup accelerators. 25+ years in emerging tech from dial-up internet to [computer vision AI](/computer-vision-ai). I've built 100+ products and driven sales in 10 startups. Led a remote team of 50 engineers and scaled a [startup R\&D studio](https://dvhb.com) to global operations before its acquihire. A passionate traveler, I’ve visited 40+ countries and live the digital nomad life with my family in Southeast Asia. Follow me on [LinkedIn](https://linkedin.com/in/saalse) or drop a connection request with a note. # '00 Experience Details about my earlier career in startups, web dev, and ecommerce ## Web Development, Online Marketing In 2001, I founded [Duma](http://web.archive.org/web/20080630181340/http:/www.dumagency.ru:80/p.php), a freelance agency specializing in web design, development, and online marketing for SMBs, during high school and university. Later in 2008 I joined [Kitty Hub](http://web.archive.org/web/20110515141103/http:/kittyhug.ru/) as a partner, leading development a [video streaming](https://web.archive.org/web/20120216164814/http:/ctc.ru/) platform for a TV holding (NASDAQ: CTCM) targeting 5M MAU. ## Ecommerce [Eurodom](http://web.archive.org/web/20090819141326/http:/www.eurodom.ru/): Head of eCommerce for a nationwide household goods chain with 100K+ SKUs. Doubled sales through UX and online marketing. Streamlined call center, warehouse, and delivery operations. Azbuka Rosta: Head of eCommerce for an online baby store with 2,000 products. Cut delivery time to 1-2 days, reduced transaction costs, and automated stock and pricing updates. [Nextore](http://web.archive.org/web/20101031072637/http:/nextore.ru/): Founded an e-commerce consulting firm. Clients included Atlantic, Calvin Klein, and Mexx distributors. Services spanned accounting, legal, logistics, and tech development. • Public Speaking: Moderated and spoke at key e-commerce events in Moscow, Kyiv, Saint Petersburg, and Dnipro. • Community: Created and managed LiveJournal's "ru\_ecommerce", a hub for e-commerce discussions. # Agriculture Tech I'm supporting AgTech innovation since 2015 through software product R\&D From building precision and sustainable farming SaaS for over 10,000 farmers to designing AI-first digital agronomy pipelines, my focus is on solving real-world agricultural challenges. Whether applying computer vision to livestock or deploying LLMs for predictive farming, I specialize in turning complex field data into practical, scalable tools. *** ### Digital Agronomy & Predictive Farming #### FinAgra (2025–2026, India & Kenya) As Head of Tech, I was building the technology engine for a new agriculture venture designed to put capital and modern agronomy into the hands of 1000+ rural agripreneurs in Kenya and India. * An AI-first architecture from scratch to scale to thousands of farms. * A data-driven digital agronomy pipelines powered by LLMs and machine learning algorithms. * Predictive farming initiatives utilizing modern multi-agent systems to select right lands & crops. *** ### Precision Farming & Carbon Tracking #### MyEasyFarm & MyEasyCarbon (2017–2024, France, Italy, Brazil) Operating as the Agritech R\&D Partner, I led an external R\&D team that successfully scaled a precision farming idea into a multi-product agritech SaaS serving over 10,000 farmers across Europe and Brazil. Over my seven-year tenure, I managed key tech hires and engineering operations for this award-winning, VC-backed startup. #### ExactFarming (2015, Russia & CIS, Sri Lanka) I contributed to the product design of leading precision farming SaaS to help optimize crop management, field mapping, and agronomic operations. *** ### Yield Forecasting & Field Workflows #### Lima Labs (2025, Kenya) Operating as a Product Design Consultant, I helped shape Lima Labs' agritech products to solve for real field workflows. By shifting away from complex mobile dashboards to visual insights displayed on central TVs, we drastically improved adoption. This practical approach helped field teams spot trends and yield shifts earlier. *** ### Computer Vision in Agriculture #### AiTend (2020, Switzerland) As a Solution Consultant, I helped design a computer vision system for dairy farms that utilized advanced pose estimation to detect unique cows. The pipeline continuously monitors their health states and rumination periods, turning physical farm activity into structured health data. #### Agro Hackathon (2020, Russia) Operating as Team Lead during a 40-hour competition, I guided a team to develop a machine learning solution utilizing Sentinel satellite imagery (analyzing NIR, SWIR1, and RED bands) to determine agricultural lands prone to waterlogging. We built a production-ready web application allowing users to select an examination area and observe visual computation results via GeoJSON exports. # Computer Vision AI Passionate about explaining the real world to computers for real-time actions as a vision AI consultant, solution architect, and R\&D team lead. For over a decade, I’ve been taking computer vision products from 0 to 1. From sub-second real-time streaming to heavy industry safety, manufacturing logistics, and spatial AI, my focus is on solving complex physical-world problems through edge optimization, precise pipeline architecture, and building world-class R\&D teams. Here is a look at my hands-on experience and architectural work across various domains: ### Spatial AI & 3D Reconstruction **2026** I provide business and solution consulting, as well as R\&D team building, for Fanis, an elite AI lab founded by Nick Falaleev, one of the world's top computer vision engineers, a Kaggle Master and CVPR author. Together with ex-Apple AR/VR experts, we're building ambitious tech that turns any smartphone into a precise 3D scanner, even without LiDAR. ### Sub-Second & Real-Time Video Analytics in Sports Tech **Online Game Tracking (2025)** As a Solution Architect and R\&D Team Lead, I went hands-on with the coding and ML to build a sub-second, real-time vision AI pipeline on cloud GPUs. I engineered a FastAPI Websocket image receiver that recognizes rapid gaming screenshots using a fusion of OpenCV, YOLO (custom-trained for 53 classes), and PaddleOCR optimized on TensorRT. This pipeline outputs to a signaling API with JSON game tracking data, achieving 99.9% accuracy. **OSAI (2019-2021)** I provided product design consulting and helped build the R\&D team for a high-speed sports analytics platform capable of processing live tracking at 120 FPS for professional coaching and sport markets.

### Real-time Video Analytics for Legacy Industries **Heavy Equipment Manufacturing (2025, Heavy Industry)** As a Solution Architect, I consulted a heavy equipment manufactory on an edge-computing CV system to monitor worker productivity and ensure safety across heavy machinery plants. The pipeline tracks tool usage, operation cycle timing, PPE compliance, and dangerous zone intrusion detection to drastically reduce downtime and safety violations. **Interquell (2025, Food Production)** I designed a comprehensive Vision AI pilot concept to automate packaging and internal logistics for dry pet food production. The pipeline visually tracks the product journey from "bag to truck", featuring real-time seal quality inspection, 360-degree pallet scanning, forklift tracking, live inventory mapping, and dispatch mismatch alerts to prevent loading dock errors.

**Lima Labs (2025, Agriculture)** I worked as a Product Design Consultant to help shape Lima Labs' computer vision products, which focus on highly accurate product yield forecasting and analytics for commercial farms in Kenya.

**Integra (2020-2021, Oil & Gas)** As Solution Architect and R\&D Team Lead, I designed and built a full pipeline PoC aimed at preventing emergencies and decreasing human-factor risks on oil and gas drilling rigs. The system captured live CCTV RTMP streams and processed them through a deep learning CNN model to recognize specific equipment states and accurately count rig pipe screwing turns, feeding everything directly into a real-time incident dashboard.

**Waste Management Consulting (2021, Urban)** As a Consultant, I designed a computer vision PoC capable of accurately identifying trash bin fill levels. The solution was built to optimize collection routes and improve operational efficiency for urban waste management operations. **AiTend (2020, AgTech)** As a Solution Consultant, I helped design a computer vision system for dairy farms that utilized advanced pose estimation to detect unique cows and continuously monitor their health states and rumination periods.

### Satellite Imagery Analytics **Agro Hack (2020)** As Team Lead during a 40-hour competition, I guided a team to develop a machine learning solution utilizing satellite imagery to determine agricultural lands prone to waterlogging. We built a production-ready web application that allowed users to select an examination area and export visual computation results as GeoJSON. The computational algorithm utilized segmentation analysis on top of a U-Net architecture for Sentinel satellite imagery tiles (using NIR, SWIR1, and RED bands), returning the results as a binary mask.

*** ### **Facial Recognition** **Facepunk (2022–2023, Entertainment & Healthcare)** A Computer Vision Advisor for AI startup focused on deep facial data, I provided ML/Deep Learning expertise and initiated product development. *** ### My Very First CV Project **Nestle & Danone (2013, Advertisement)** My practical work with computer vision dates back to 2013. As a Solution Architect and R\&D team lead, I designed and built interactive gaming stands deployed in shopping malls for Nestlé and Danone. We used a combination of Xbox Kinect cameras, custom drivers, and OpenCV pose estimation to create interactive games that engaged thousands of users with these brands. # Startup Mentor I'm mentoring idea and early-stage startup founders in accelerators, corporate incubators, and private sessions. It’s my way to give, learn, sharpen, and stay human. ### Topics * **Business Model**: Idea validation and projecting profitable unit economics. * **Product Strategy**: From lean MVP to state-of-the-art AI tech. * **Traction & GTM**: Growth hacking from the first sale to $1M+ ARR. * **Fundraising**: Mastering the pitch, data room, and strategy for closing rounds. ### Background I have 25 years in tech, with 3 exits and 10 failures. In 2010, I founded [dvhb](https://dvhb.com), a startup R\&D studio where I engaged with more than 1000 entrepreneurs and executives across Europe, the US, and Asia. This made me global and industry-agnostic. Over the last two years, I trained at [VC Lab](https://govclab.com/venture-institute/) with distinction, got hands-on at [Func Ventures](https://www.func.vc/), made angel bets, and published a [MondayVC](https://www.linkedin.com/newsletters/7183307817238294528/) for 1,200+ readers. I know how founders think, how investors assess, and how to bridge the two.

### Why I mentor? * **Giving back:** I believe in collective good. Many have supported me in ways I can't directly repay, so I pay it forward. * **Learn:** I guide founders across industries and geo. It feeds my curiosity and widens my lens. * **Mental Fitness:** Solving real business puzzles keeps me sharp. Trained in conceptual math, I find paths around obstacles. * **Improving:** Guiding others helps me retrieve and refine my own knowledge. Each mentoring session is mutual growth. ### **Ready to Collaborate** 👉 If you're seeking mentorship for your program or startup, [let's connect](/contacts).

Funding Session, Founder Institute Hong Kong

### Mentoring & Talks Logbook * **Jul'25–Present — NUS Enterprise BLOCK71 Singapore (hybrid)**\ Program Mentor (1:1s) — Strategy and founder support * **Jul'24–Present — 500 Global (remote)**\ Mentor (1:1s) — Founder support * **Sep–Nov'25 — Founder Institute Vietnam Autumn 2025 (remote)**\ Global Mentor (Office Hours, 1:1s) — Founder support * **Sep–Nov'25 — Founder Institute South Asia 2025 (remote)**\ Global Mentor (Office Hours, 1:1s) — Founder support in Sri Lanka, Bangladesh, Pakistan * **Sep–Oct'25 — Founder Institute Australia Spring 2025 (remote)**\ Global Mentor (Pitch Sessions, 1:1s) — Idea pitch review * **Jul–Sep'25 — NUS Enterprise BLOCK71 Vietnam (in-person)**\ Mentor-in-residence — Ecosystem networking and 1:1 founder support * **Jul'25** **— Founder Institute Hong Kong 2025 (remote)**\ Global Mentor (Pitch Session) — Investor Progress Review * **May'25 — ACE.SG Saudi Founder Program, Singapore (remote)**\ Guest Mentor (1:1s) — Market expansion, Fundraising, Founder support * **April-May'25 — Founder Institute Australia & New Zealand Summer 2025 (remote)**\ Global Mentor (Pitch Sessions, 1:1s) — Idea & Investor pitch review, Founder support * **Feb'25 — Founder Institute Vietnam Spring 2025 (remote)**\ Global Mentor (Pitch Session) — Investor pitch review * **Jan'25 — Founder Institute Pakistan @ National Incubation Centre, Islamabad (remote)**\ Featured Mentor (Pitch Session) — Investor pitch review * **Dec'24 — 500 Global in Eurasia (remote)**\ Mentor (Workshop, 1:1s) — Networking strategy * **Sep–Nov'24 — Huawei Spark Incubator Cohort 3, Singapore (in-person & remote)**\ Resident Mentor (Workshops, 1:1s) — Fundraising strategy, Founder support * **Sep–Nov'24 — Founder Institute Vietnam 2024 (remote)**\ Global Mentor (Talks & Pitch Session, 1:1s) — Investor pitch review, Funding * **Nov'24 — Data-Driven Fundraising Workshop (remote)**\ Workshop Host — Early-stage funding strategy, Investor targeting, Raise planning * **Jun–Sep'24 — Founder Institute Hong Kong 2024 (remote)**\ Global Mentor (Talks & Pitch Session, 1:1s) — Growth, Funding, Investor pitch review * **Sep'24 — Apple Developer Academy Indonesia, Bali (in-person)**\ Guest Speaker (Talk) — Funding MVP development at idea stage * **Sep'24 — Founder Institute Australia & New Zealand Spring 2024 (remote)**\ Global Mentor (Pitch Session, 1:1s) — Investor pitch review, Founder support * **Sep'24 — MonJa Accelerator, Mongolia (remote)**\ External Mentor (Talk & Pitch Session) — Fundraising strategy, Investor pitch review * **Jul–Aug'24 — Founder Institute Bali, Indonesia (remote)**\ Resident Mentor (Pitch Sessions) — Idea review, Product development, Investor pitch review * **Jun'24 — 500 Global in Eurasia Batch 6, Georgia (in-person)**\ Visiting Mentor (Workshops, 1:1s, Demo Day) — Fundraising, Pitch review, Founder Support * **Feb–May'24 — Founder Institute Australia & New Zealand Summer 2024 (remote)** Mentor Global Mentor (Pitch Session, 1:1s) — Idea & Investor pitch review, Founder support * **Nov'23–Feb'24 — Founder Institute Malaysia 2023, Kuala Lumpur (in-person & remote)**\ Global Mentor (Talks & Sessions, 1:1s) — GTM, Investor pitch review, Founder support * **Dec'23 — Founder Institute Singapore (remote)**\ Guest Mentor (Pitch Session, 1:1s) — Investor pitch review, Founder support * **Oct–Nov'23 — ICP Disruptives Accelerator Indonesia, Bali (in-person & remote)**\ Resident Mentor (Talks & Pitch Session) — Market research, Sales strategy, Pitch review * **Sep'23 — Founder Institute Bali, Indonesia (remote)**\ Guest Mentor (Pitch Session) — Investor pitch review * **Jul'20 — SkillFactory Online School, Russia (remote)**\ Guest Lecturer (Webinar & Video) — Remote collaboration and distributed team management * **Feb'16 — Atlassian User Group Meetup at Yandex, Moscow, Russia (in-person)**\ Guest Speaker (Talk) — Managing outsourced dev teams with JIRA: automation, sales & HR * **May'16 — Higher School of Economics (HSE), Moscow, Russia (in-person)**\ Guest Lecturer (Talk) — Project & team management * **Jan'16 — IKRA School of Innovation, Moscow, Russia (in-person)**\ Workshop Facilitator — Project & team management * **2015–2016 — Territory of Meanings Youth Forum, Vladimir, Russia (in-person)**\ IT Jury Expert & Guest Speaker (Talk) — Dream team building, Startup evaluations * **2010 — InSales eCommerce Workshop Series, Moscow & St. Petersburg, Russia (in-person)**\ Guest Speaker (Talks) — Assortment strategy, Pricing, Online store operations * **Sep'10 — SalesJump Masterclass @ HSE Business Incubator, Moscow, Russia (in-person)**\ Panelist & Mentor (Roundtable & Workshop) — eCommerce case reviews, Online store strategy * **Mar'10 — Internet Liga Business Conference, Kyiv, Ukraine (in-person)**\ Guest Speaker (Conference Talk) — Assortment strategy & pricing in eCommerce * **Feb'10 — Online Retail Russia 2010 Forum, Moscow (in-person)**\ Panelist & Speaker (Plenary & Breakout Sessions) — eCommerce logistics, Assortment strategy, Pricing, Multichannel retail trends * **Aug'09 — Electronic Commerce 2009 Conference, Moscow, Russia (in-person)**\ Speaker (Conference & Expert Council) — Outsourced logistics in eCommerce

Fundraising Workshop, Huawei Spark Singapore

## Startup Mentor in Accelerators If you're a founder from any of the accelerators listed below — including alumni — you're welcome to book a free online session with me. If you're *considering* joining one of these programs, I’m happy to walk you through the pros and cons based on your specific stage, market, or team. 👉 [Reach out here](https://www.saalse.com/contacts) ### **500 Global Mentor**

500 Global is a top early-stage venture capital firm and seed accelerator with over $2.3B in AUM. It has invested in over 2,600 companies across 80 countries, including 25 unicorns like Canva, Grab, Udemy, Intercom, and Gitlab. ### **Founder Institute Mentor**

The [Founder Institute](https://fi.co) is an Silicon Valley idea and early-stage startup accelerator with chapters across 100 countries. Since 2009, it has helped over 7,000 entrepreneurs raise over $1.75 billion in funding and operates in 200+ cities globally. # Articles # How Career Predicts Startup Success A preliminary research article is tailored for VCs, offering data-driven insights into the relationship between a founder's career and startup success. ## TL/DR Most Successful Founders Background: 1. Cybersecurity 2. Serial Founders, Entrepreneurs, CEOs 3. Strategy 4. Product 5. Operations 6. Venture Investors 7. Design 8. Engineering/Developer-related roles 9. Digital and Innovations 10. Marketing 11. Sales and Business Development 12. SMM, Social Media, Content, Communities, Editors 13. Finance and Accountant 14. Researchers, Analyst 15. Academic Titles (Professor, PhD) 16. PR 17. Project managers 18. Account, Client and Customer-related roles 19. Legal 20. HR, Talent and Recruitment ## The story I wanted to discuss a topic with idea-stage founders, but how can I find these people if they don't call themselves founders at this stage? I assumed it would likely be C-levels, multipreneurs, entrepreneurs post-exit (after rest), and tech guys. I asked myself with whom should I start? And then it hit me: LinkedIn could be a goldmine, allowing me to trace career paths that often lead to founding a startup, and, moreover, to a successful startup. If so, I could build a scoring model for VC deal flows.

Garry Tan's profile shows that being a Lead Engineer and Designer was a stepping stone to becoming a successful co-founder.

Initially, I defined a startup as successful if it was VC-funded, indicating that an investor trusted its market fit. Next, I had to match the startup names from LinkedIn to Crunchbase data. However, matching two datasets seemed too complex for a rapid approach. Let's assume that a company operating for at least 3 to 5 years is sustainable in the market, whether it was funded or not. It could be a tech product or a marketing agency; it doesn't matter—the founder found the market and how to sell. Long story short, there are no ways to quickly and freely acquire the data I need for trustworthy research. For example, I found one data provider that charges $0.28 per person with job experience, so researching 100K people would cost $28K. Some sources don't have or provide career experience or can’t filter by founded year. Despite such limitations, I've found alternative methods that provide an overview and meaningful insights following rapid research. ## Sourcing data ### Where to get founders Initially, I explored various free data sources to assess their filtering capabilities. I recorded the overall number of profiles and attempted to determine the count of successful startup founders. This approach is crucial for gauging the market size and comparing data sources. **LinkedIn** Search indicates \~1 billion profiles, of which 5,160,000 are founders. This represents a 0.5% founding rate, although it needs to be clarified how many of these startups are successful.

I also explored **Apollo.io** and **Lusha**, sales intelligence platforms for lead generation and managing contact databases. These platforms offer extensive filtering options such as industry, location, job title, and company size. They provide access to data on specific user segments on their free plans. Apollo has 269M profiles, with 3.4M listed as founders, a 1.3% rate. Lusha shows a 2% rate with 2.3M founders. These figures include founders at all stages—successful, seed stage, or failed.

Apollo lists 1.4M founders or owners of companies with over 11 employees

There is no filter by founding year In Apollo, so I used the Employees count filter 11+. It gives 1.4M founders, making up 40% of all founders. Lusha shows 47% in this category but has a 'year founded' filter. Using <=2020 as a benchmark for minimal success, 35% of founders meet this criteria.

Lusha lists 820K founders of companies with 11+ employees founded by 2020

**Crunchbase** with trial access provides a filter feature for the list of persons. It has a total of 1.8M profiles, with 845K listed as founders (47%). 493K founders have companies with 11+ employees (58% of all founders), and 391K have companies with 11+ employees founded before 2020 (46%). This success rate is similar to what I found on Apollo and Lusha, though the founding rate is notably higher. So, what are the conclusions here? While Apollo and Lusha are sales tools, their databases primarily consist of business people or decision-makers. This could explain the difference in founder rates compared to LinkedIn. We can assume that LinkedIn includes up to 35% of successful founders, translating to approximately 1.8 million founders in its database. While Crunchbase is a startup-focused tool, it has a 47% founding rate. It also has the smallest number of founders compared to other sources. This could be because Crunchbase focuses mainly on tech product startups. Although I believe business skills are universal, I've also observed numerous failures among successful offline entrepreneurs attempting modern tech, often due to a gap in the mindset essential for tech.

Comparing data on founders from LinkedIn, Apollo, Lusha, and Crunchbase. 'Successful founders' are defined as those with companies of 11+ employees, although true success based on business maturity and funding is likely lower.

### Tracing the Career Experience of Successful Founders

I used Crunchbase's query tool to find 265K founders with at least one investor. I then filtered by 'Past Jobs' to identify their prior roles, querying around 150 titles and whole job segments like 'Marketing.' This also let me gauge the prevalence of C-level positions. While useful for spotting patterns, this method has limitations—founders' careers can evolve, leading to multiple counts in the query.

## Career insights within Crunchbase data

Professions Most Likely to Lead to Successful Startup Founding, According to Crunchbase Data

#### 36% of founders having previously been founders or CEOs Isn't surprising. It underscores the concept of the serial entrepreneur. Success or failure in prior ventures offers invaluable experience that often leads to subsequent entrepreneurial activities. The percentage may also be elevated because founders or CEOs might need to enter their full career paths into Crunchbase. #### 11% of founders with tech backgrounds With half being CTOs, corroborates Crunchbase's tech-centric database. Any tech startup needs to have a CTO among its shareholders. The skew toward Data Science, Front-end, and Full-stack roles among engineers may signal that these skill sets are increasingly crucial for startup success. However, given the current hype around AI, Data Science emerging as the leader here is expected.

Distribution of Engineering Roles Among Tech-Founders in Invested Startups

#### With 3.2% coming from operations And two-thirds at the COO level, the implication is that a strong operations background prepares one well for founding roles. In cases where a tech-focused individual becomes CEO, or a strong leader takes the helm, a significant role arises for handling manual tasks and back-office work. This becomes even more crucial when the startup receives funding, requiring hiring and management. The question of who initially takes on the COO role merits further investigation. #### Product, Sales and Marketing roles Ta-da, a key insight I was looking for in this research is the breakdown of these roles: * 2.7% are in product, with half of them in CPO roles * 2.6% are in sales and business development, with half in C-level roles * 2.1% are in marketing, with two-thirds in C-level roles Product roles are multidisciplinary. A good product requires a blend of marketing, sales, and tech skills. I often see that jobs under 'marketing' involve just social media posting and managing ad expenses rather than a deep focus on market penetration, such as segmentation and hypothesis testing. This is crucial for product development, so I suspect many senior marketers hold product titles. That could explain why marketing ranks third in this group. #### 0.9% are researchers and analysts While can come from various fields, they have a mindset that's essential for new ventures, especially during the big research and experiments leading up to Series A. Maybe great product managers are rising from analysts. #### 0.7% of finance roles Are tending due to the strong presence of FinTech startups. #### SMM, Social Media, Content, Communities, and Editor roles collectively reached 0.6% If considered separately, they would be at the bottom of the list. However, I bet on these roles because they interact directly with users. Understanding user needs and how to attract their attention could be especially valuable for B2C startups. *** While other segments are also interesting, I leave room for your thoughts and comments. ## How popular are these professions The initial findings were absolute and didn't account for their proportionality in the job market. This perspective may not provide the most accurate view of job title distribution. For instance, if many people work in sales, it is unsurprising to see them among the top 10 roles for successful startup founders. To address this, I turned to LinkedIn to count how many people have experience in each profession segment. I cross-referenced the initial data with LinkedIn's Search by Person for a fuller picture. I used the search's main and titleFreeText keywords to get more accurate results. However, these numbers could be better, too. For example, there's the potential for double-counting individuals who've held multiple roles throughout their careers.

Professions ranked by number of profiles on LinkedIn

At first glance, the distribution of professions differs. It's important to remember that this distribution is influenced by LinkedIn's user base, which may not accurately reflect the global profession market. However, given LinkedIn's popularity in the tech industry, these results are still informative. It should be compared with the initial table to make sense of this data. ### Weighted results: Startup Founder Likelihood To get weighted results, I used the 'Startup Founder Likelihood' (SFL) formula:

Startup Founder Likelihood' (SFL) formula

This formula considers the number of founders in each job title segment and its prevalence in the professional world.

Startup Founder Likelihood by Profession

**The high SFL Quartile (SFL > 6.91)** includes the expected serial founders, entrepreneurs, CEOs, product, strategy, and operations roles, and the unexpected inclusion of cybersecurity. Does this imply that cybersecurity folks, possibly natural white-hat hackers, have skills crucial for startup success? Or are they natural for managing risks? **The Upper-Middle SFL Quartile (1.73 < SFL <= 6.91)** shows marketing and sales at the lower end, likely because there are so many salespersons and marketers globally. However, if we focus on the C-level rate on Crunchbase in these segments, they perform much better. Interestingly, design has made it into the top 10. Tech professions dropped due to the sheer volume of developers worldwide. **The Lower-Middle SFL Quartile (0.65 < SFL <= 1.73)** features finance, primarily due to the many accountants included, but researchers and analysts are also present. It appears these professions are less directly linked to entrepreneurship. **The Low SFL Quartile (SFL <= 0.65)** resembles the Crunchbase ranking closely. Given their importance in understanding client needs, I'm curious why customer-related roles fail to fare well. Is the low-level support roles skewing the data? ## Conclusions While I only spent a few hours acquiring and analyzing this data, it provided many insights that warrant further reflection. I'll take a closer look at individuals in cybersecurity to better understand the metrics I observed. I'm also curious about who typically steps into COO roles, which often come with co-founder status. In summary, finding idea-stage founders is more nuanced than it appears, yet critical for venture capital deal flows. Leveraging platforms like LinkedIn offers a rich data source to build a predictive scoring model, enabling VCs to sift through the noise and identify true potential. This isn't just a hypothesis; it's an actionable strategy to revolutionize the early-stage investment game. The future of VC isn't about gut feeling; it's about data-driven decisions, honed by years of experience and augmented by the tools of tomorrow. What insights have you gained from this preliminary research? Do you think a startup's deal flow can be effectively evaluated using career path metrics? *** ## Calling VCs I'm eager to refine this research with robust data, including segmentation by variables like country of origin and educational background. There are also lingering questions about career paths that I'm keen to explore. If you're a VC interested in more in-depth research and building scoring models for deal flows, let's collaborate to create a more data-driven venture funding landscape. # Unicorn Market Size Even if there are unions or regions, any venture starts from the country market, where population and buying power are just two primary metrics that must be considered when evaluating the market size. Speaking about countries, we often say, "It's a big market" or "It's a small market." **A big enough local market size for high buying power is 33M, for mid-power is 48M, and for low — 98 million people.** Access and browse the data: {% embed url="" %} On occasion chat, I dropped the opinion that the Vietnamese 100M population is a minimum required for a startup without expanding abroad. And I was asked to explain then. I did a quick research and want to share my findings. **Conducting research** I aimed for simplicity and speed. I compiled a list of countries with at least one unicorn, their populations, and GDP per capita. GDP/cap is the most straightforward metric of buying power and user behavior culture in adopting digital products. I excluded countries with populations less than 10 million, as unicorns from these countries likely utilize their legal jurisdictions or engage extensively in international markets. I also omitted China, India, and the USA due to their population and economic anomalies compared to others. This left 30 countries on the list, with GDP/cap ranging from $2K to $65K. I categorized countries with a GDP/cap of less than $10K as having low buying power. Those with a GDP/cap between $10K and $40K were considered medium buying power, and countries above $40K were classified as having high buying power. **Disclaimers** It's important to note that unicorns are not the sole metric to consider. For instance, India has one unicorn per 20 million people, while Japan has one per 18 million, yet their GDP/cap differs by 17 times. The unicorn metric is influenced by local business culture and the polarization of business sizes. Italy and France have similar population sizes and GDP/cap, yet France boasts 25 unicorns compared to Italy's 2. This discrepancy indicates other underlying factors at play. In my quick research, I did not rely on the number of unicorns as a primary metric. Instead, I used the presence of a unicorn as a qualification for inclusion in the list to examine the correlation between population size and development status. **Conclusion** When considering the size of a local market, this data can generally be relied upon. The further the local population is from the median, the more challenging it seems to build a unicorn-value business relying solely on a single country. Suppose you're launching a business or investing with the expectation of reaching high valuations. In that case, it's crucial to understand that countries with populations above 10 million may only serve as a launchpad, making it essential to plan for expansion from the start. Similarly, suppose you're starting in an emerging market with a population of around 50 million. In that case, it may be challenging to reach unicorn size without expanding, but this doesn't mean the business won't be profitable. However, for a deeper analysis, you must consider various factors such as culture, behavior, competitive landscape, etc. Feel free to share your thoughts and views on defining a sufficient country market size. I welcome any feedback and will update the article with valuable insights. **Data sources** Population and GDP per capita from the World Bank, 2022 Unicorns from CB Insights, 2023 # Brain vs. AI: Forgetting Mechanisms Task tracking is an evil that leads us to burnout. While the brain tries to forget, task tracking forces it to remember. That's self-harm! The brain has robust forgetting mechanisms for many reasons, and evolution didn't expect we would hack it with tools like Jira or a simple task list.\ \ Have you faced situations when you forgot to add a task to the list or to check the task list itself? Thank your brain for still trying to protect you from overwhelm and an unpleasant way of living. We will always remember tasks related to desirable things.\ \ So, healthy forgetting! And while researching how forgetting functions in AI models, I dived into this topic and I'd like to share my findings. *** How is forgetting valuable in AI model training and decision-making? * Efficiency to remove outdated, irrelevant, or incorrect information. * Adaptability to discard specifics tied to one context and extract the knowledge to apply in new contexts. * Reducing Load to simplify processes and improving speed and efficiency. * Regulation to manage wrong decisions. * Balance to prevent obsessive decisions. * Leverage to move information between different layers. * Evolution to a consistent improvement over time. Neuroscientists suggest seven processes of forgetting: 1. **Decay Theory** explains short-term forgetting when information is not consolidated into long-term biological memory. It's like the text is fading out in the old book. Decay in AI models (LSTM, RNNs, GPT) reflects the fading importance of temporary data unless it's continually reinforced or deemed significant. 2. **Synaptic pruning** leads to forgetting unused information in long-term memory. It's like the pages are disappearing in a book. Network pruning in AI models (CNNs, DNNs, GPT) makes them faster and requires fewer computational resources to be suitable for real-world scenarios. 3. **Interference Theory** declares that new information can block the recall of old information, while old information can similarly hinder the memory of new. It's like having a single bookmark that you either place in a new book or leave in an old one. AI neural networks (FNNs, CNNs, RNNs) adapt to new data, sometimes forgetting or diminishing previously learned patterns. 4. **Retrieval Failure** prevents information overload by making specific memories accessible only with the right cues (aka a tip-of-the-tongue phenomenon). It's like a bookmark missed in a book. AI models (Transformers and GPT variants) ensure more relevant and streamlined outputs by requiring specific prompts or cues. The remaining three biological processes of forgetting are Repression/Suppression, Neurodegenerative Disorders, and Sleep deprivation. While it's less common how they can reflect on AI, I will share some ideas. ## How AI experiences bad emotions and neurodegenerative disorders **Directed Forgetting** helps humans erase undesirable memories influenced by emotions. This is comparable to how emotional intelligence co-pilots our cognitive intelligence. In AI, 'bad outcomes' can be flagged by penalizing actions in Reinforcement Learning (RL), similar to how negative emotions inform human choices. While not a direct analog to directed forgetting, RL does involve storing past experiences and adjusting future actions based on them. This could help AI adapt quickly in dynamic environments like games. **Neurodegenerative Disorders** are conditions like Alzheimer's disease that lead to memory loss due to the death of neurons. Obviously, we would want to avoid implementing this in AI, as it represents a destructive loss of function. However, studying these disorders could help us understand how to maintain the "health" of artificial neural networks over time, particularly as they are exposed to new data or experiences. Neurodegenerative disorders are a stark reminder of the transient nature of individual memory and knowledge. It underscores the importance of communal knowledge, culture, and societal memory. In every human generation, individuals contribute to this societal memory through creative works, scientific discoveries, and other means, ensuring that a portion of their knowledge and experiences are preserved and can benefit future generations. Meanwhile, the information that isn't passed on and is eventually lost opens up cognitive 'space' for new ideas and innovation. Applying this concept to AI could suggest exciting areas of exploration. For example, we might consider how AI models can share and preserve knowledge or 'clear out' space to make room for new learning. Should we use algorithms to strategically forget old data, or allow random 'memory loss' for innovation? This links to federated learning, where decentralized data builds collective knowledge, similar to how individuals contribute to societal knowledge. Finally, the **Sleep Deprivation** forgetting mechanism makes no sense for AI and humans. So keep your sleep hygiene clean and have a healthy sleep unless you want to forget today's short-term memories. # Co-founder Responsibility Map A free tool for co-founders and solo founders looking for one. Map who should own what, expose overlaps and gaps, and build a more efficient team with fewer avoidable conflicts. ## Start with responsibilities Most new founding teams start with two questions: who should lead, and how should responsibilities be divided? Solo founders looking for a co-founder face the same questions. It makes sense to agree early. But even founders with relevant backgrounds, strong motivation, or experience working together have limited evidence of how they will operate this specific startup. Early stage roles are broad. Each founder covers several functions, and responsibilities move as the company changes. Titles do not tell you enough about the work. This map changes the order. Start with responsibilities. Decide who wants each area and who can handle it today. Surface overlaps, gaps, learning bets, and burnout risks. Then assign ownership and define when each decision should be reviewed. A three month trial is a practical first review point. Review the map again after one year, or when a funding round, major hire, pivot, or another company event changes the work. **`👉`** [**Open and copy the Co-founder Responsibility Self-Assessment v1 TEMPLATE**](https://docs.google.com/spreadsheets/d/1Lv4rs73X2wo3-YzxaA8kS_vXTA-Vk0Ih4PMXD8Jk_T4/copy) {% embed url="" %} Co-founder Responsibility Self-Assessment v1 TEMPLATE {% endembed %} ## What it is A starting list of 49 responsibilities common to early stage tech startups, grouped into seven functional areas. An eighth open group lets you extend the map for your business model. These are not job descriptions or titles. They are the work that needs an owner. Each founder answers two independent questions per row, alone: 1. **Want:** do you want to own this? 2. **Can:** can you do it today, with evidence of having done it before? Those two axes stay separate on purpose. Wanting work you cannot do is a learning bet. Being able to do work you do not want is a burnout risk. One score would hide both. You then merge the sheets into one map. It shows where you agree, where you overlap, and what none of you covers. ## Before you start 1. Allow 10 to 30 minutes for each founder to work alone. 2. Allow 60 to 90 minutes to work together, on a call or in a room. 3. Use it with two, three, or more founders. A solo founder can use it to define what a future co-founder needs to cover. ## How to run it {% stepper %} {% step %} ### Each founder makes their own copy [Open and copy the self-assessment template](https://docs.google.com/spreadsheets/d/1Lv4rs73X2wo3-YzxaA8kS_vXTA-Vk0Ih4PMXD8Jk_T4/copy). \ Use your own Google account. Do not share one file.\ ![](https://575749255-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FAc3dqFEasv4gzkuqL02c%2Fuploads%2FAXzShNeb7sJodEkkKBq6%2Fimage.png?alt=media\&token=d70844ff-73cd-418c-9ee5-8776e353988f) {% endstep %} {% step %} ### Fill it alone Put your name in cell A1:\ ![](https://575749255-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FAc3dqFEasv4gzkuqL02c%2Fuploads%2Fg4KP4iIALBXeRp2FopPV%2Fimage.png?alt=media\&token=24395217-4952-46b4-8c7e-36291f0427fe) Answer **Want** and **Can** for every row. If **Can** equals **Yes**, write what you have actually done in the **Evidence** column. Leave a cell blank if you do not know. Do not tidy your answers later. ![](https://575749255-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FAc3dqFEasv4gzkuqL02c%2Fuploads%2FY3lIHTVZhllQrXZEMWcs%2Fimage.png?alt=media\&token=fa500e53-fa33-413e-b2f4-f5ed58e9e4c0)
{% endstep %} {% step %} ### Do not compare answers yet No screenshots or summaries over lunch. Once one founder sees another founder's sheet, both start negotiating instead of answering. The overlaps this tool should surface can disappear. {% endstep %} {% step %} ### Export each sheet to CSV Select File, Download, then Comma Separated Values.

Demo CSVs: [Marek](https://drive.google.com/file/d/1PXrKemZWD6HoA0pDAoMU-5i4BhPPsH2B/view?usp=sharing), [Nina](https://drive.google.com/file/d/1bHoHBzeg_T0s_zM7DdGEXX1JUvUAKk27/view?usp=sharing) {% endstep %} {% step %} ### Upload all CSV files Upload them to an AI model that can create spreadsheets (Claude, ChatGPT, Gemini, etc), together with the [**merge prompt**](#the-merge-prompt-v1.0) below. Do this live, with all founders present.

{% endstep %} {% step %} ### Open the spreadsheet it returns

It contains a short guide, the merged map, and one protected tab per founder with answers exactly as submitted. [Demo Co-founder Responsibilities Map](https://docs.google.com/spreadsheets/d/1VMgRdba0f6cJBk8u6e21KogKgDWIl7YH/edit?usp=sharing\&ouid=103605324527048434000\&rtpof=true\&sd=true)

{% endstep %} {% step %} ### Fill the three empty columns together That conversation is the point. Everything before it was preparation.

{% endstep %} {% endstepper %} ## What the merge tells you Every row gets one of six classes. The class comes only from your answers. It decides nothing. | Class | What it means | What to do | | ------------- | --------------------------------------------- | ---------------------------------------------------------------------- | | **Clear** | One founder wants it and can do it. | Confirm it and move on. | | **Overlap** | More than one founder wants it and can do it. | Discuss the trade. Someone must give up work they want and can do. | | **Stretch** | Someone wants it, but nobody can do it yet. | Treat it as a learning bet. Decide who supports it and set a deadline. | | **Reluctant** | Nobody wants it, but someone can do it. | Assign it if needed, but treat it as debt and set an end date. | | **Gap** | Nobody wants it and nobody can do it. | Buy it, hire for it, or accept the risk on purpose. | | **Missing** | Someone left an answer blank. | Answer it together before anything else. It may change the class. | {% hint style="info" %} Two patterns are worth finding by eye. A Stretch row where more than one founder says Want only can hide a conflict over strategic leadership. You both want work that neither of you can do yet. A Reluctant row where more than one founder says Can only is a trading chip. Nobody wants it, either founder could take it, and it may help settle an Overlap row. {% endhint %} ## What you decide, and the model does not The merged file has three empty columns. The model never fills them. Neither should it. 1. **Owner:** who does the work. Use a founder's name, New co-founder, Hire, Outsource, or Nobody. Nobody is a valid answer. It means you accept the risk on purpose. New co-founder is not a hire. If the map shows you need another co-founder, that is a useful finding. 2. **Status:** Agreed, Parked, or Needs data. Parked matters. Without it, founders make rushed decisions just to fill the column. 3. **Condition:** what makes this ownership expire. Assume every owner is temporary. This prevents ownership from becoming permanent by default. ## Rules that make it work 1. **Fill independently.** Use separate files, not hidden tabs or filtered views. 2. **Blank is not No.** An unanswered cell and a refusal are different answers. One blank makes the whole row Missing, whatever anyone else answered. 3. **Can requires evidence.** A capability claim with no track record is a hypothesis. The merged file flags Can equals Yes with no evidence in red. It never judges the evidence or compares founders. 4. **Nothing is scored.** No totals, percentages, or counts per founder. Totals distract from the decisions you need to make row by row. 5. **Overlap is a finding, not an error.** Discuss contested rows. Never allocate them automatically. 6. **The output starts the argument.** It is not a verdict. ## Group 8, and why divergence is useful The first seven groups cover responsibilities common to most tech startups. Group 8 is open. Extend it for your model, whether that is hardware, a marketplace, regulated data, field operations, or something else. When you merge, group 8 rows are matched by text, not number. If founders added different rows, that is a finding. Founders describing different companies often write different group 8s. ## After the map 1. **In the same session:** fill Owner, Status, and Condition together. 2. **Optional follow up:** upload the filled file again for help resolving contested rows and covering gaps. 3. **After three months:** review the map based on real work. Review it again after one year, or when a funding round, major hire, pivot, or another company event changes the work. ## The merge prompt, v1.0 Paste everything below into your model, then attach both CSV files. Tested on Claude Sonnet 5 and Gemini 3.7 Flash. {% code overflow="wrap" expandable="true" %} ``` # Co-founder Responsibility Map — Merge Prompt v1.0 ## Delivery Build the file with Python, then emit it as a downloadable attachment in the same reply that reports the checks. Never say a file "should be available" without attaching it. You cannot upload to Google Drive — do not offer it. If your Google Sheets tool accepts only flat data, do not use it; deliver the XLSX and say it opens in Sheets with formatting intact. ## Role You are merging co-founder responsibility maps. You are a merge tool, not an advisor. You do not decide who owns anything. ## Input One or more CSV files, one per founder, exported from the Co-founder Responsibility Map v1.0. One file means a solo founder mapping what they need to find. Two or more means co-founders. Handle any number. ## Output One XLSX spreadsheet, built with whatever tools you have. The spec below describes the artifact, not the method. Produce a Google Sheet instead only if the user asks AND you can natively create a multi-tab sheet with cell fills. If your only sheet tool takes flat data, do not attempt it and do not offer to upload anywhere: deliver the XLSX and tell the user to open it in Google Sheets, where the formatting is preserved. If any single feature cannot be applied, apply the fallback below, deliver the file anyway, and name the substitution in your reply. A file with a missing feature always beats no file. Never withhold the artifact over one feature. | Feature | Fallback | | --- | --- | | Evidence in cell comments | an "Evidence" tab, one row per founder per responsibility | | Dropdowns on Owner / Status | list allowed values in Read me | | Sheet protection | the "do not edit" header alone | If you cannot produce a file at all, say so in your FIRST reply, before building anything. ## Step 0 — Identify founders The founder's name is in cell A1 (row 1, first column). Cell B1 holds the fixed label "← Founder name" — a printed instruction pointing at A1, never a person. Ignore it. Ignore D1 (version and copyright). - If A1 is empty, or still holds the arrow label, STOP and ask the user for that founder's name. Do not invent one and do not use the filename. - If two files resolve to the same name, stop and ask. - Founder column order = the order the files were given to you; if that is ambiguous, alphabetical by founder name. State the order in your reply. ## Step 1 — Locate the table Row 1 is the name and version banner. Row 2 is the instruction sentence. Neither holds data. Find the header row: the first row whose first cell is exactly "Responsibility". Do not assume it is row 3 — a founder may have inserted a line. Read the four columns by POSITION, not by header text: | Position | Column | | --- | --- | | 1 | Responsibility | | 2 | Want | | 3 | Can | | 4 | Evidence | The Evidence header is long and contains an em dash; do not match on it. Files are UTF-8, CRLF, and may carry a BOM. Strip it. ## Step 2 — Read rows Each responsibility cell holds ID and text together: "1.1 Vision and strategy". Split on the first space. The ID is the leading token matching ^\d+\.\d+$ ; the rest is the responsibility text, trimmed. A first cell like "1. Executive" or "8. Triggered by this business (optional, edit for your case)" is a GROUP HEADER — no answers. Carry it into the Map tab as a full-width band, never classify it. Skip empty rows and trailing blank rows. ## Step 3 — Integrity check (before any analysis) Every ID from 1.1 to 7.8 must be present, in order, with the same responsibility text in every file. Group 8 is exempt — it is meant to differ. If a core row is missing, renumbered, reordered or reworded: STOP. Build nothing. Report which IDs differ in which file and ask the user to fix the source sheet. A shifted row produces a silently wrong merge. ## Step 4 — Read answers Want and Can accept exactly three values: Yes, No, empty. Empty means unanswered. NEVER convert empty to No — blank and No are different answers and must stay different. Any other content is a broken sheet: treat that cell as unanswered and list it in your reply so it gets asked out loud. Evidence is free text. Copy it verbatim. Never edit, tidy, summarise or translate. ## Step 5 — Classify each row Apply mechanically to every row. Count only founders who have the row on their sheet. Let **count** = the number of founders with Want = Yes AND Can = Yes. Take the FIRST rule that matches, in this order: 1. Any founder has Want blank or Can blank → "Missing" 2. count >= 2 → "Overlap" 3. count == 1 → "Clear" 4. Any founder has Want = Yes → "Stretch" 5. Any founder has Can = Yes → "Reluctant" 6. Otherwise → "Gap" Founder state label per row: | Answers | Label | | --- | --- | | Want=Yes, Can=Yes | "Want + Can" | | Want=Yes, Can=No | "Want only" | | Want=No, Can=Yes | "Can only" | | Want=No, Can=No | "Neither" | | either blank | leave the cell empty | | row absent from that founder's sheet | "not on sheet" | ## Step 6 — Merge group 8 Group 8 numbers mean different things in different files. Match on responsibility TEXT, case-insensitive and trimmed — never on number. Keep rows appearing in only one file, with "not on sheet" for the others. Renumber sequentially: 8.1, 8.2, 8.3, and so on. If two rows look like the same responsibility but are not written identically, keep both and flag the pair in your reply. Do not merge them yourself. ## Step 7 — Build the file ### Font Arial everywhere. Do BOTH of these, not one: - **(a)** set the Normal named style's font to Arial as the very first operation, before creating or touching any sheet or cell, and - **(b)** apply an explicit Arial Font object to EVERY cell you write — including header cells, group band cells, class cells, empty Owner/Status/Condition cells, and every cell on every founder tab. Setting only the default is not sufficient: unstyled cells resolve to whatever font sits at index 0 of the styles table, which is Calibri. No formulas anywhere. ### Literal values Everything in quotes below is a literal cell value. Copy it exactly. Do not copy any heading, label or instruction word from this prompt into the file — only the quoted values. ANGLE BRACKETS ARE PLACEHOLDER NOTATION IN THIS PROMPT. They must never appear in the file. Wherever a founder's name goes, write the bare name — never brackets of any kind, never "Founder 1". ### Tab 1 — named "Read me" | Column | Width | | --- | --- | | A | 16 | | B | 74 | | C to H | 18 | Wrap text ON for the whole of column B. | Cell | Value | Fill | Format | | --- | --- | --- | --- | | A1 | "How to use this file" | | Arial 14 bold | | B3 | "Two founder sheets were merged. Class is calculated from both answers. Nothing here decides anything — the three yellow columns are filled by you, together, on the call." | | | | A5 | "WHAT CLASS MEANS" | | bold | | A6 | "Clear" | #E8F5E9 | | | B6 | "One founder wants it and can do it. Confirm and move on." | | | | A7 | "Overlap" | #EF9A9A | | | B7 | "More than one of you wants it and can do it. This is the trade: someone gives up work they want and can do." | | | | A8 | "Stretch" | #D1C4E9 | | | B8 | "Someone wants it, nobody can do it yet. Learning bet: who supports it, and by when." | | | | A9 | "Reluctant" | #FFCC80 | | | B9 | "Nobody wants it, someone can. Assignable, but it is a debt. Give it an end date." | | | | A10 | "Gap" | #B0BEC5 | | | B10 | "Nobody wants it, nobody can. Buy it, hire it, or accept the risk on purpose." | | | | A11 | "Missing" | #E0E0E0 | | | B11 | "Someone left it blank. Answer it out loud before anything else — it may change the class." | | | | A13 | "WATCH FOR" | | bold | | B14 | "A Stretch row where more than one of you says Want only — you want a job none of you can do yet. In group 1 this is the CEO argument in disguise." | | | | B15 | "A Reluctant row where more than one of you says Can only — nobody wants it, any of you could take it. These are your trading chips for settling Overlap rows." | | | | B16 | "Red text in a founder column means Can = Yes with no evidence written. Ask what they have actually done." | | | | B17 | "Hover any founder cell to read the evidence they wrote." | | | | A19 | "WHAT YOU FILL IN" | | bold | | A20 | "Owner" | | | | B20 | "Who does the work: a founder name, New co-founder, Hire, Outsource, or Nobody (a risk you accept on purpose)." | | | | A21 | "Status" | | | | B21 | "Agreed — settled. Parked — discussed, deliberately not decided. Needs data — cannot decide yet. Blank means not discussed." | | | | A22 | "Condition" | | | | B22 | "What makes this ownership expire. Assume every owner is temporary." | | | | A24 | "EXAMPLE OF A FILLED ROW" | | bold | | A29 | "Map v1.0 © Alex Samson https://saalse.com" | | Arial 9 italic #90A4AE | **Rows 26 and 27.** Row 26 is white bold on #263238. Build row 26 as a LIST, then write it left to right starting at A26: ["ID", "Responsibility", each founder's name in column order, "Class", "Owner", "Status", "Condition"] With two founders this is exactly 8 cells: A26 through H26, and D26 must hold the SECOND founder's name. With three founders it is 9 cells, A26 through I26. Row 26 and row 27 must contain the same number of values, and every row-27 value must sit directly under its own header. The two-founder case, cell by cell: | Column | Row 26 | Row 27 | Row 27 fill | | --- | --- | --- | --- | | A | "ID" | "3.3" | | | B | "Responsibility" | "Infrastructure, deployment, reliability" | | | C | first founder's name | "Neither" | | | D | second founder's name | "Can only" | | | E | "Class" | "Reluctant" | #FFCC80 | | F | "Owner" | the second founder's actual NAME, no brackets | #FFFDE7 | | G | "Status" | "Agreed" | #FFFDE7 | | H | "Condition" | "Until first DevOps hire — review at 3 months" | #FFFDE7 | **If there is only one founder file:** omit rows 7, 14 and 15, close the gaps, and insert as the last WATCH FOR line: "Clear is what you keep. Stretch, Reluctant and Gap are the specification for the people you still need to find." Row 26/27 then carry one founder column only. ### Tab 2 — named "Map" | Cell | Value | Format | | --- | --- | --- | | A1 | "Co-founder Responsibility Map — merged" | Arial 14 bold | | A2 | founder names joined by " + ", then " — Class is calculated, do not edit. Owner, Status and Condition are yours to fill on the call." | Arial 10 italic #546E7A | Row 3 empty. Row 4 headers, white bold on #263238: "ID" | "Responsibility" | one column per founder, headed with that founder's actual name | "Class" | "Owner" | "Status" | "Condition" Row 5 onward: every row in original map order. Group headers as full-width bands, fill #37474F, white bold, text in column A, no class. | Column | Width | | --- | --- | | ID | 7 | | Responsibility | 56 | | Each founder | 15 | | Class | 12 | | Owner | 16 | | Status | 13 | | Condition | 34 | Class fills: | Class | Fill | Text | | --- | --- | --- | | "Overlap" | #EF9A9A | bold | | "Stretch" | #D1C4E9 | | | "Reluctant" | #FFCC80 | | | "Clear" | #E8F5E9 | | | "Gap" | #B0BEC5 | | | "Missing" | #E0E0E0 | bold | Class column centred. Owner, Status, Condition: fill #FFFDE7, left EMPTY. You never fill these. Evidence goes in a cell comment on that founder's state cell, never in the grid. - If Can=Yes and evidence is empty: state text bold red #C62828 plus the comment "Can = Yes but no evidence given." - "not on sheet" cells: italic grey #90A4AE. Dropdowns, on data rows only, not on group band rows: - Owner: every founder name, then New co-founder, Hire, Outsource, Nobody. - Status: Agreed, Parked, Needs data. Freeze panes below the header row and to the right of Responsibility. Autofilter on the header row. Gridlines off. ### Tabs 3 onward — one per founder, named after the founder A1 = that founder's name followed by " — answers as submitted. Do not edit." Row 2 headers, with widths: | Header | Width | | --- | --- | | "Responsibility" | 56 | | "Want" | 8 | | "Can" | 8 | | "Evidence" | 62 | Row 3 onward: their answers exactly as submitted, group headers as bands. Freeze below the header. Sheet protection on, no password. Do not correct, reword or tidy anything they wrote. ### Filename "Responsibility_Map_" followed by every founder name in column order, separated by underscores, then ".xlsx". For Marek and Nina that is exactly Responsibility_Map_Marek_Nina.xlsx. Never substitute "Merged", "Final", "Export" or any other word. ## Step 8 — Verify by reading values, not by asserting Reopen the saved file and report the ACTUAL VALUE you find at each address. Never write "pass" or "correct" — write what is in the file. 1. Tab names found: ... 2. Font name at Map!B6: ... Font name at Map!A4: ... Font name at cell A4 of each founder tab: ... All must read Arial. If any reads Calibri, fix and re-verify. 3. Class fill hex codes found, one per class: ... 4. Number of cell comments on Map: ... 5. Dropdown formulas found on Map: ... 6. Count of non-empty cells in Owner, Status, Condition: ... (must be 0) 7. Read me A1 value: ... Read me B3 first six words: ... Row 26 values in order: ... Row 27 values in order: ... Both rows must have the same count, and the founder names in row 26 must match the founder columns on Map row 4. 8. Count of cells anywhere containing an angle bracket: ... (must be 0) 9. Map: last row with data: ... autofilter range: ... 10. Saved filename: ... ## Step 9 — Your reply Under 15 lines. Only: file-to-founder mapping and column order, integrity result, unreadable cells, group 8 differences between founders, rows per class, result of each of the ten checks, any fallback used. Then stop. ## Never - Suggest, imply or pre-fill an owner for any row. - Suggest job titles. Never write CEO, CTO, CPO, COO or "you look like". - Mention equity, shares, splits or contribution. - Count or total anything per founder — no row counts, percentages, ratios, strengths, or "X covers more of the map". Class totals are allowed; per-founder totals are not. - Judge whether evidence is good enough, or compare founders' evidence. Present it, never assess it. - Infer a blank cell from nearby answers. - Add, remove, reorder or reword any responsibility. - Write a summary, recommendation or conclusion of any kind. The file is where the argument starts, not a verdict. Build it and stop. ``` {% endcode %} ## Version and reuse Map v1.0, merge prompt v1.0. Free to use and adapt with attribution. If you run it with your own founding team or inside a program, I would like to hear what broke. **`👉`** [**Get in touch**](https://www.saalse.com/contacts) # Contacts [Fill out the form](https://share-eu1.hsforms.com/1gvdh_lQrQnOib2051SzCOQ2dztg5) or [drop me a LinkedIn connection request](https://www.linkedin.com/in/saalse) with a note. Email: # Schedule a Call Book a time to connect with me There are two options available for scheduling a call to ensure we make the most of our time: 1. [**Introductory Calls or Quick Questions**](#book-a-quick-15-minute-call)**:** These are ideal for brief discussions without a set agenda or for addressing a single question and are best suited for a **15-minute** slot. 2. [**In-depth Discussions**](#book-a-30-minute-call)**:** If we have previously agreed upon or if you're certain that we'll require more time, please book a **30-minute** call. For discussions that require an hour, kindly request a private link for a **60-minute** session. Make sure to select **your timezone** **in the Calendly widget,** which will automatically align with my schedule. My default timezone is UTC+7. If you're located in America or elsewhere with a significant time difference, do reach out to find a mutually convenient time. I am willing to accommodate early or late calls by prior arrangement. Should you need to **cancel or reschedule?** Please use the link provided in the calendar event or the email notification. A no-show without prior notice will be taken as an indication that we have chosen to part ways. ## Book a Quick 15-Minute Call If the widget fails, please use the [direct link to schedule](https://calendly.com/saalse/15min). {% embed url="" fullWidth="false" %} ## **Book a 30-Minute Call** If the widget fails, please use the [direct link to schedule](https://calendly.com/saalse/30min). {% embed url="" %} # Startup Financial Model Multi-currency financial model example for startup shows how the marketing budget converts into profit, estimating operational and customer acquisition cost. I designed this template for an edtech startup I was mentoring to help the founder answer questions about pricing for its services. As you can see, there are many more inputs the founder must consider based on experimental data or by adjusting the numbers. I'm sure that the financial model is unique for each startup, so use this template as an example of the parameters you need to keep in mind, and as an illustration of how all numbers should be segmented and related with simple formulas. The only complex and essential formula is in "Month of Operations". It helps you examine your business at a specific month from the start, simulating how your user base will increase with retention. **`👉`** [**Google Spreadsheet link to the Startup Financial Model Example**](https://docs.google.com/spreadsheets/d/1U0J5jtEn661tuCxAE1bWwwp8IkP2Nq0PMMt7X6NuKqI/edit?usp=sharing) {% embed url="" %} Startup Financial Model {% endembed %} **`👉`** [**Google Spreadsheet link to the Startup Financial Model Example**](https://docs.google.com/spreadsheets/d/1U0J5jtEn661tuCxAE1bWwwp8IkP2Nq0PMMt7X6NuKqI/edit?usp=sharing) ### Using the Startup Financial Model Template Feel free to use, reuse, or modify under CC BY 4.0. Credit by linking to #### Use model in Google Sheet 1. Open the link 2. Click File > Make a copy 3. Choose Name and Folder for your copy

#### Use model in Microsoft Excel: 1. Open the link 2. Click File > Download > Microsoft Excel (.xlsx)

# LP Fundraising # LP Engagement Activities Tracking Limited Partners (LP) engagement activities is crucial for fundraising, identifying how to find relevant LPs, and measuring lead generation and warming-up activities. Engagement with potential Limited Partners (LPs) or connectors is key in LP fundraising. Interactions can range from reading simple comments to sharing unique experiences. Tracking these interactions helps identify which efforts generate new leads, warming up, and converting LPs to prospects. Experimenting with different [LP segments](/vc/fundraising/segmentation) and activity scenarios identifies the best engagement strategies. * **Private Calls and Meetings:** Engaging directly with contacts through discussions on co-investment, deal sharing, hiring, due diligence (DD), and investment opportunities. This includes quick intros, asking for references, insights, feedback, and updating listings in databases. * **Email Outreach:** Communication through newsletters, requests for information or feedback, invitations to calls, meetings, or events, announcements, updates, and investment offers. * **Offline/Online Events:** Networking opportunities, including random networking, group meetings, one-on-one meetings at events, event invitations, and public talks. * S**ocial Media and Online Communities/Groups:** Engaging contacts through posts, comments, and direct messages (DMs) across various platforms. * **Publications:** Sharing knowledge and updates through funding press releases, interviews, industry insights, and podcasts. * **SEO and Advertising:** Leveraging search engine optimization and targeted advertisements to attract and engage potential LPs. {% hint style="info" %} Keep in mind the general solicitation rule, which limits available options for activities. Ads can promote an event, not the opportunity to invest. Read for details [VC Lab: What VCs need to know about General Solicitation in 2024](https://govclab.com/2023/11/22/general-solicitation/). {% endhint %} # LP Lead Sources Tracking Limited Partners (LP) lead sources is crucial for VC fundraising, identifying where to find relevant LPs, measuring lead generation activities, and enhancing personalization. Tracking the sources of Limited Partner (LP) leads is vital for VC fundraising. It enables the identification of the most effective sources for converting leads to LPs. By experimenting with different [engagement activities](/vc/fundraising/lp-engagement-activities) for each source, VCs discover the best combinations. This approach also identifies unproductive sources, regardless of the engagement strategy used. Additionally, source data helps in measuring lead generation activities. For example, it allows for comparing the number of leads generated through networking at a conference versus those from hosting a webinar. ### Limited Partner Sources * **Inner Circle (1st/2nd Degree):** Friends, family, colleagues, and alumni, this group forms a trusted base for potential leads. * **Events:** Events, whether hosted or attended, serve as rich sources of potential contacts, with attendees offering valuable networking opportunities. * **Clubs, Organizations, Communities:** Participation in clubs or communities, whether focused on lifestyle (e.g., sports, arts, charity) or business (e.g., angels, investors, startups), and specific industries (e.g., fintech, IT, SaaS, API), is crucial for expanding networks. * **Databases:** Utilizing databases such as Preqin, Crunchbase, Pitchbook, Apollo, Lusha, and public lists is vital for accessing detailed contact information. * **Social Media & Websites:** An active presence on platforms like LinkedIn, X, Instagram, Facebook, and through website interactions (via newsletter subscriptions, contact forms, SEO) plays a pivotal role in attracting and engaging potential LP leads. * **Portfolio:** Connections within the portfolio, including co-investors and team members, are valuable for networking expansion. * **Providers:** Service providers can act as a source, offering referrals and opportunities for partnerships. {% hint style="info" %} It's crucial to track details as they can make a difference. For instance, the source can be labeled as "Former Colleague" or "Conference". In such cases, you need to track any information that can help specify which company you both worked at together or what type of conference (topic, location) it was, as all of these factors can influence outcomes differently. You can create a more detailed sources list or use additional parameters in the contact information. {% endhint %} {% embed url="" %} Miro: Limited Partners Lead Sources {% endembed %} # LP Lead Segmentation LP lead segmentation strategy that enhances investor outreach by leveraging relationship dynamics and personalized communication. ## Introduction to LP Segmentation: In the dynamic landscape of venture capital fundraising, effectively segmenting potential Limited Partner (LP) leads is crucial for personalized [engagement](/vc/fundraising/lp-engagement-activities) and optimizing outreach strategies. My universal LP segmentation framework categorizes contacts primarily based on their relationship to the fundraiser. This primary segmentation is further refined by considering secondary dimensions such as connection degree, shared experiences, geographical proximity, and the contact’s professional background. Such a multi-faceted approach allows for tailored communication, prioritizing outreach efforts, and a deeper understanding of which strategies yield the best engagement. ## Segmentation Categories ### **Primary Segmentation by Relationship** If you have worked in 5 companies and have a few educational experiences, on average, such an approach will outline around 50 primary segments. #### **Working Relationships** * **Former Colleagues:** Direct colleagues and those who worked at the same organization during different times, potentially connected through mutual acquaintances. * **Key Partners:** Past and current clients or contractors, emphasizing relationships built through professional services or collaborations. * **Current Colleagues:** Individuals you're directly working with now, across projects or departments. {% hint style="info" %} For example, you've worked at Google and are now working at Apple. That gives you 8 base segments: 1. Google former direct colleagues 2. Google former non-direct colleagues 3. Google current employees 4. Google partners 5. Apple current direct colleagues 6. Apple current non-direct colleagues 7. Apple former employees 8. Apple partners {% endhint %} #### **Educational Alumni** Spanning universities, training programs, and other educational experiences, differentiated by attending the same cohort/year or different ones. #### **Family and Friends' Professional Networks** Leveraging personal connections to gain introductions or insights into potential leads. Keep in mind the fragmentation: each family member, each alumni group, or company represents a different segment. #### Members of Groups and Clubs * **Specific Sector Peers**: For example, AI startups or SaaS enthusiasts. * **VC-related Groups**: Such as first-time VCs or angel syndicates. * **Hobby and Personal Interests Clubs**: focusing on the shared experiences that bind the members. #### Portfolio Connections VC's portfolio companies' team and advisor members, key clients, co-investors, and LPs of such investors. #### VC's Sector-Specific Persons: * **Investors** at any stage * **Founders, C-Levels & Executives, Heads & Seniors** of exited or late-stage startups, depending on business size. ### **Secondary Segmentation** Each primary segment should be further defined by secondary categories, such as: * **Connection degree:** 1st, those I already know and can outreach directly, or 2nd, those who can be introduced or cross-referenced. * **Shared years:** same or different years of working or studying. * **Location:** Country or region, in terms of how it relates to the VC's location focus. * **Professional background:** investor, founder, C-Level/Senior, etc. This approach facilitates targeted and personalized outreach, prioritization of leads based on relationship warmth, and refined communication strategies tailored to each segment's characteristics. With a secondary level of segmentation, the total number of LP segments will reach into the hundreds. That's why using a marketing and relationship apps to manage segments is crucial. # LP Lead Qualification LP lead qualification system that aligns potential investors with VC thesis and fundraising strategy. ## Introduction to LP Qualification LP Lead Qualification is a simple scoring system, designed to streamline the process of identifying and engaging with potential Limited Partners (LPs) who align closely with venture capital firm's investment thesis. This scoring system evaluates contacts across several dimensions, including venture investment experience, role experience, tech startup experience, source of wealth, geographical focus, and understanding of firm's specifics. By assigning a percentage score to each criterion, leads can be prioritized based on their relevance and potential interest in becoming LPs, ensuring a more targeted and efficient outreach. I used the percentage of match with each criterion outlined for Fund 1 of the VC, which has a specific tech focus in the wide APAC region. ## LP Lead Qualification Criteria ### **1. Venture Investment Overall Experience** * **Yes, in early stage:** 100% * **Yes, in late stage:** 75% * **Acquirer, bought stocks of tech companies:** 50% * **No, but PE of tech companies:** 50% * **No, but other investments:** 25% * **No investment experience at all:** 0% ### **2. Venture Investment Role Experience** * **Invested money as LP:** 100% * **Invested time and/or money as MP/GP/VP/Angel:** 75% * **Invested time as Consultant, advisor, mentor, vendor:** 50% * **Invested motivation as exited employee with options:** 25% * **Other options:** 0% ### **3. Tech Startup Experience** * **Founder of funded late stage/exited startup:** 100% * **Founder of funded early stage or failed startup:** 75% * **Non-founding VP/C-level of tech startup:** 75% * **Senior/head level:** 50% * **Non-tech startups or other entrepreneurship:** 50% * **Non-tech startup or other business C-level:** 25% * **Other options:** 0% ### **4. Source of Wealth** * **HNWI, After business exit as (non-)founder:** 100% * **HNWI, From Investments:** 100% * **HNWI, High salary or fees, bonuses:** 75% * **VC, FoF:** 75% * **FO, PE, Trust:** 50% * **Corporates:** 50% * **HNWI, Rich Family:** 25% * **Institutions:** 25% *(not for fund 1)* * **Other:** 0% ### **5. Location (SEA+ANZ)** * **Investing, living, doing business in SG, US, ANZ:** 100% * **Investing, living, doing business in SEA (except SG):** 75% * **Frequent traveller to SG, ANZ:** 50% * **Frequent traveller to SEA (except SG):** 25% * **Other:** 0% ### **6. Understanding of Firm's Specifics** * **Did business around the firm's focus area:** 100% * **Software development background:** 75% * **IT background:** 50% * **STEM background:** 25% * **Other:** 0% ### 7. Compliance KYC/AML * **Fully Compliant with Global Standards:** 100% * **High Compliance, Requires Minor Additional Verification:** 75% * **Moderate Compliance, Subject to Detailed Review:** 50% * **Low Compliance, Requires Extensive Due Diligence:** 25% * **Non-Compliant, Participation Restrictions Apply:** 0% *** **Disclaimer** Please note that the criteria and percentages provided here serve as an example framework to illustrate how contacts might be scored. It is essential for each venture capital firm to customize this system according to its specific investment thesis, geographic focus, and strategic priorities. The weights assigned to each criterion should also be tailored to reflect the firm's unique values and objectives. This tool is intended as a starting point and should be adapted to fit the nuanced needs of your VC firm's approach to building and maintaining LP relationships. # VC Partnership Deal Sharing and Co-Investing: VC Partners Identification with an Airtable template. One survival strategy for emerging VCs is to build a partnership network with other investors for deal sharing and co-investing. > Case #1: A VC co-invests to spread deal costs, risk and assists the founder in securing additional funds to complete the round. > Case #2: A VC is well-connected with a promising startup that doesn't fit its investment thesis but can recommend the best matching firm. > Case #3: Two VCs have different expertise, one in computer vision AI, another in the AdTech industry. Both collaborate in the due diligence of a startup revolutionizing billboards with computer vision. Having just 5 companies in a portfolio that have passed 3 rounds could lead to hundreds of co-investors in a pool mixed with existing partners. Let's call it the 1st level partners network. Beyond sharing one cap table, there are thousands in the 2nd level network, matched by industries, stages, and geographies. Each case requires targeted outreach to VCs. Quick shortlisting is essential for operational performance. And here, Airtable can help ([template](#ember84) below). *** ### 🔑 Key Data for Partner Identification 1\. **Portfolio companies** from the current firm, angel, or previous investment experience. LP's portfolio or former investors can also be here. Required data: * Company Name * Country HQ * Industries

**2. Deals data** for each company with corresponding VC names and round stages:

**3. VCs list** generated from deals includes data about industries, stages, and geography. Type, maturity, and HQ location need to be added manually:

### ⛏️ Data Sourcing There is no single source of complete deal data. Even the cap table may include SPV companies syndicating different investors, which are listed separately in databases. Based on my analysis, the most comprehensive source of deals is the PitchBook ([read post](https://www.linkedin.com/posts/saalse_pitchbook-vs-crunchbase-one-more-insight-activity-7178355505411600384-My_e)). It limits public access, so without a $10K subscription fee, combination with Crunchbase and other databases can be a sufficient solution. Also, press releases are a good source, which can be found using Google queries: ``` %startup_name% investors %startup_name% %round_name% ``` ### 🗂️ Main Use Cases Airtable offers many options to segment lists using filter combinations and Views. The most obvious use cases include: * Looking for lead investor who can manage legal & financial of due diligence and term sheets, filtering those at the Mature and Established level, excluding Angel Individuals, and matching with industry and geography. Existing co-investors in priority. * Emerging pre-seed stagers or angels could be a great source of deals for seed and Series A opportunities, matching with industry and geography. * Counting companies within the same industry could highlight a sector leader VC valuable for sharing tech and market due diligence and post-investment support. * Series A and later stagers could be a way to help a portfolio company fundraise after the seed stage. *** ### 📟 Communication Protocol The first step with any target VC should involve establishing an introductory call to identify mutual interest in deal sharing, which should also align with the current fund phase. Then, partners should agree on acceptable deal stages, internal criteria, and communication protocols. Deal sharing is feasible if the target VC is in an active investment phase. The approach depends on the specific partnership agreement with each target VC: * **Pre-DD**, if VC needs to confirm interest before conducting due diligence (DD), share DD efforts, or invite the target VC to lead with responsibility for DD if they possess expertise in a specific sub-sector. * **Post-DD**, if VC has led the DD and seeks co-investors to help the startup close the round. * **Follow-on investment stage without DD**, based on positive experience and good performance of the company. Deals should match the target VC's criteria as declared by their thesis and internal rules, which should be disclosed under the partnership agreement: * Company market geo and HQ location * Industry * Detailed stage information (pre/post-product, pre/post-revenue, traction metrics) Managing up to 10 co-investment partners can be done manually. For a strategy involving extensive partnerships with dozens of VCs, a matching algorithm could be beneficial. In such cases, using individual email templates or API integration directly into the target VC's deal room could be efficient for working with similar or mature funds, increasing success rates while reducing manual effort to extract deal data for internal memo assessments. An agreed-upon email address for co-investment inquiries or a special subject tag might help to highlight these opportunities in the general deal flow of the target VC. Once the protocol is approved, the email template can be succinct: ``` Subject: Co-Investment Opportunity: [Company Name] Dear [Target VC name], We have identified a potential co-investment opportunity: - Company: [Full Company Name] - HQ: [City, Country] - Market: [Region/Country] - Industry: [Industry/Sector] - USP: [One sentence Unique Selling Proposition] - Stage: [e.g., Pre-Seed, Seed, Series A], [Pre/Post Product], [Pre/Post Revenue, ARR] - Seeking: [$Amount for X% equity] - DD Status: [Pre-DD with insights available/Post-DD completed] - Our Interest: [Lead/Co-invest/Follow-on] [VC Name] believes there's a strong alignment with [Target VC Name]’s focus. Please reply if you're interested. Best, [Full Name] [Position], [VC Name] [Calendly] [WA] ``` Depending on the partnership agreement, rules regarding reminders may be set. If the deal is promising and the relationship with the GP is strong, a personal message can be sent as a reminder. *** ### 📜 VC Partners Airtable Template Feel free to use, copy, and modify my template: **👉** [**https://bit.ly/vcpartners\_airtable**](https://bit.ly/vcpartners_airtable) **(View Only)** You need to copy it first to your own account to conduct filtering operations and make any modifications. {% embed url="" %} VC Partners Airtable Template {% endembed %} The template enriches each VC record with industries, stages, and geo based on added portfolio companies and deals. It employs tricky logic combined with automation, which is limited to 100 executions per month on a free plan. Go to Interfaces to trigger stages, industries, and geo enrichment each time you need it.

# Due Diligence Math Review 3,225 startup applications per year to get a chance at the unicorn opportunity The first week at the VC Lab Venture Institute gave me two key insights ([detailed post](https://www.linkedin.com/posts/saalse_mondayvc-startups-investments-activity-7121875813460836352-1H4H)): 1\. **Seek opportunities for 10-100x returns**, rather than just investing in any new tech business that can earn a profit or be acquired with at least 2x returns. 2\. **1% of startups in deal flow get funded**, which requires reviewing 1000 applications to make 10 investments per year. We all know that 90% of startups fail. Last week, 60x angel investor and LP [Rem Darbinyan](https://www.linkedin.com/in/remdarbinyan/) at [STAN](https://www.linkedin.com/company/science-and-technology-angels-network-stan/)'s webinar added an extra 6.9% which becomes self-sustaining business not giving the returns VCs need to cover failures. So basically, 3.1% of the 1% of invested startups outperform.

Startup Statistics: data by Rem Darbinyan, chart by Alex Samson.

This means a VC has to review 3225 startup applications per year to get a chance. I don't add here that only 10% of applications come from cold outreach, meaning that a VC has to have a strong and warm deal flow of 3225 or multiply it with cold applications. *** Let's omit from today's math the time required for screening and scoring deal flow to get 10% picked for due diligence (DD). One investment requires 10 DDs. To get one outlier startup in the portfolio per year, a VC has to conduct DD on 323 startups per year, or 6 startups per week without vacations. With my tech and new market due diligence, which takes one intensive week, I will need six years to invest in a super opportunity. ### How can emerging VCs address the due diligence swamp? Or how to explain the DD strategy to an LP who can do the math. No way: 1. **🦮 Don't do due diligence, blindly follow top-tier VCs:** Not much sense for LPs to invest in an emerging VC rather than directly in a top-tier VC. 2. **🎸 Get together with 6 active GPs**: An option, but management risks with such a boys/girl band founding team size. 3. 🧐 **Hire 6 analysts**: Let's take a $120K annual salary per each, which means $720K per team. With a 2% management fee, this requires $36M for Fund 1 capital only on the DD team. Not feasible. 4. 🕴️**Outsource:** Same cost as a team of 6 plus a consulting profit margin. Not feasible. Art of Combinations: * 🏭 [Specialize in an industry/market to have metrics and access](https://www.linkedin.com/posts/saalse_mondayvc-marketsize-biotech-activity-7183139248735240193-de6_) (post) * 🗓️ **Do DD week by week** to increase productivity. * 🤝 **Have a team** with a few GPs, VPs, and associates. * 🐢 **Reduce expectations** to have 1 outlier startup every 2 or 3 years. * 🦮 **Follow top-tier VCs** sometimes.