How to Analyse Startup Ideas: Our Plan for 12 Startups in 12 Months

Mar 14, 2020

My business partner and I were due to be in San Francisco next week to start the Deep Learning Part 1 course at University of SF taught by Jeremy Howard. Unfortunately due to COVID-19, this course has been cancelled.

As such we will still being doing the course but manually ourselves via the free course contents online. If anyone else wants to collaborate with us get in touch.

Much like dealsourcr, we plan to launch many highly profitable startups over the next 12 months. Our target is 12, but it's likely to be less.

This was inspired much by Jon Fook and @levelsio, 12 Startups in 12 Months. This is a great method to experiment and try as many ideas as possible.

Look, you can love your ideas/businesses all you like but what matters is spending your time on the highest ROI activities. We found at dealsourcr, that just simply innovating was not increasing growth.

I think this is massive flaw in the current startup mindset, some products simply need a large distribution partner or investment to grow. Organic viral growth is absolutely possible but only on certain products/ideas.

As such, we have agreed a distribution deal with a large company. More news once the deal is signed.

After consuming every bit of content we could find, it was clear that spending some actual time upfront would save us 6-12 months of pain. For example, the last 9 months at dealsourcr could have been compressed into 3 if we had done the analysis ahead of time.

A great talk on this is from the founder of HomeJoy at YC, I highly recommend you watch it.

We are dumb chimps that learned to talk with many flaws in our thinking, so we need a rigorous approach to stop our biases influencing us...

This came partly from Jon Fooks article, a bunch of YC lectures, Warren Buffett & Vinod Khosla.

The way we think about it is as follows:

• (Potential Upside x Probability of Success) ÷ Time Investment

For example, if you can build a potential POC in an afternoon and upside is huge, it's worth doing this. But building a POC for 1 month+ that going to make you £200/month is pointless. It's really surpising, how many ideas don't stack.

But it's also not that simple, some ideas are massive, with no business model but easy to monetise at-scale. Anyway here's the scoring...

1. Hair on Fire (is there a desperate need?)
2. Organic Distribution (can this grow fast organically?)
3. Possible Distribution Partner (do we have a deal agreed before writing a line of code?)
4. Little Competition (inspired by Peter Thiel's - Competition is for Losers)
5. Passion (does this really really really excite us? do we really care about the problem?)
6. Moat (aka. Buffet's Durable Competitive Advantage)
7. Low Investment MVP (can we build in 1 hour or 6 months of hell?)
8. 0 to £1B (is this a zero to £1M idea or zero to £1B idea? VINOD - 00:40)

We score these all from 0-10, then add up all the scores and rank them accordingly.

It's very surprising how terrible your ideas really are, either they require massive investment (months and months of coding / ton of cash) with low probabilty and small market or little investment but no market. Trust me, try it on your ideas, you'll be shocked.

Until the end of March 2020 we doing this on 30-40 ideas, in April we will start building these ideas. Treating them all as experiments to try out.

Much of this came from studying Vinod Khosla, who has made it clear many times that regardless of all his analysis (and high intellect), much of ventures is about having a large number of experiments. Meaning: even the ones you think will work, don't and vice-versa.

We can do our absolute best and still fail OR be barking up the wrong tree altogether. The main principle is to have a high learning rate, constantly adapt and be willing to kill our old ideas if they do not work anymore.

Thanks for reading and keep checking out the blog! We will post all the ideas here soon. Ash.