This week at SLUSH Singapore, I gave a presentation on the New Rules of Fundraising - how entrepreneurs can prepare for the new, data-driven startup ecosystem powered by AI-based engines such as HAL (YCombinator) and DART (Hatcher) . Here's the slides and some notes from the talk - and thanks to all for your positive feedback. I had a blast giving the presentation, and I hope this explanation of these new technologies will help you in your search for funding.
The first thing entrepreneurs need to be aware of is that data-driven, rule-based assessment of early-stage companies is here to stay. The traditional methods of judging funding applications are no longer scaleable, and in our view, are not democratic, fair, or a good use of resources. Using data-driven approaches, we can ensure that your business idea is judged more precisely on merit, clarity of thought, the quality of the team, and - most importantly - on whether or not the idea is suitable for VC backing.
Slide Set 1: The Odds
We've talked to a *lot* of fund managers and investors over the course of the last two years. And what is amazing about these discussions is how closely the data correlates when it comes to successfully making it through the "funnel" - and into an accelerator. It turns out that whether you're applying in Delhi, Dubai, Seoul, or Silicon Valley - the odds of being accepted into an accelerator - or funded by a VC range from roughly 0.5% to 2%, with small exceptions above that for specialist shops:
Slide Set 2: Five Rules for Getting Accepted (into an accelerator or early-stage VC)
The fives rules here would appear on first blush to be common sense - but it is surprising how many entrepreneurs neglect to take care of the basics. In the new data-driven ecosystem, it becomes essential that you take an objective look at your target market *before* submitting your plan. If the market isn't big enough, or you don't include a solid analysis of your competitive landscape, your application will score lower (sometimes much lower) than one that includes such an analysis. And as for including your best friend as co-founder... if they don't have the ability to add significantly to the chance of success, keep them as your best friend, but don't include them as a co-founder. You need a team that can execute every aspect of your business... building a good product will prove to be just a small part of success in your later years. You need strong expertise in sales, pricing, marketing, and scaling - and reporting. And you need an executable plan as to how you intend to scale.
Slide Set 3: Five Tips for Getting Funded
Getting into an accelerator is, most of the time, only half the journey. Not every graduate of your cohort is going to get funding - in fact the odds are less than 50% at most accelerators (YCombinator being the notable exception - a whopping 4 in 5 of their grads get funding - all the more reason why you need to make sure your application to YC is "HAL-ready".
Re Tip #1, many of you are no doubt familiar with the excellent work of Michael Lewis, and his book "Moneyball" - a case study in how data can be used to downgrade the importance of subjective systems of selection in favour of objective data models that focus on performance. There are some very strong similarities between how baseball scouts *used* to operate, prior to the emergence of Moneyball's data-driven techniques, and how venture capitalists sometimes choose entrepreneurs.
Note: One of the aspects of the selection process that I am *really* tired of is the tendency to look for a particular founder "type" - something I call the "Charming Founder Syndrome". You know the type - a handsome, typically male, 20-something grad from a top English-based university that has a knack for using the right kind of words, and more than enough charm to sail through difficult questions concerning the challenges his business is likely to face.
I've seen more than a few of these guys get funding on the basis of thin presentations - and more than a few "non-charming" founders get rejected. It drives me crazy, and is one of the reasons I love the fact that we're moving away from the pre-Moneyball "scouting" methodologies to data models that ignore over-large personalities to objective, data-driven scrutiny of business plans.
Summary: Data is Good. Data Can Level the Playing Field
As I write this, I'm sitting in Itaewon (a hip area close to downtown Seoul), reading Weapons of Math Destruction by Cathy O'Neil, a data scientist who writes passionately about the dangers of over-reliance on algorithm-based data models (the irony of reading a book with that title while *real* WMDs lurk a mere 28 miles away and world leaders trade kindergarten-grade insults at each other is not lost on me!)
O'Neil makes a number of very good points about the need to use robust, transparent models that incorporate large data sets and robust feedback loops. She extolls the virtues of "Moneyball" models derived from decades of data on one of the world's most popular sports, but warns against subjective analysis and systems that are biased or easily gamed.
Thankfully, there is a wealth of data available - and there is an abundance of quality feedback available from professionals in the industry that can be used to tweak the models and make them better. Yes, it's early days, but count me as one of the believers - I believe that as these models evolve, we will increasingly become able to seek out entrepreneurs on the basis of merit, and become less reliant on charming founders.
Confidence in your ability to execute comes in many forms - not everyone is able to stand up on a stage and move a crowd. What's important is your ability to move a crowd of customers - and sometimes all that takes is the power of an idea - and an ability to focus on a vision, rather than a product (see Tip #2 above.)