How do I start my new venture? / The factory playbook

“Do you have a master plan? A venture validation play book?” Asked a wise ex-colleague and reader of one of my recent posts on landing page creation.

I looked at him incredulously: “Of course”, I said.

“Why haven’t you written it down?”

Good question.

Like most of us, any coherent thoughts that come out of my mouth do so because they are part of a mental framework that I’ve built - usually on the shoulders of giants. In this case, many giants. Did the way I pieced together this particular puzzle really add anything new?

After a little introspection, it seems clear that I can hardly preach an agile, MVP mindset and hold tightly to that type of insecurity.

So, with much fanfare, here it is - the accumulation of my venture building experience over the last twenty years distilled into a step by step guide.

It’s based on the three pillars that drive my view of the world: Human centred design, lean startup methodology and agile.

The running order takes us through two supporting elements followed by 8 steps. I’d call them 8 steps to the epiphany, but you know, one of the giants already took that.

• obligatory caveat. One size does not fit all. This is simply the mental model on which my mind settles by default

It’s very early stage

A little framing first. You won’t see anything beyond the holy grail of product/market fit here.

This isn’t because I don’t think it’s interesting to scale companies, I’ve just spent most of my career working either at the seed end or in already established “startups” like eBay where the scaling part is already done - so it’s what I know.

I also hold the, perhaps misguided hypothesis that if you prime the rocket and aim it correctly at the moon, the journey in-between should be overwhelmingly dull.

What’s the goal?

So what are we trying to achieve?

Find a group of people (segment) who have a burning need (problem) that you can feasibly (economically, technically) solve (solution). Combined, these tell you everything you need to know.

That frequently quoted but ambiguously defined “product/market fit” is all about executing for the above elements. That’s the prize.

To get there you validate each of the elements in turn, which is what the 8 steps below will help you achieve.

You know you’ve hit product/market fit when you have two things:

• The elements pieced together to create, in essence a small machine (or perhaps a factory is a better analogy). You plug money into the machine and get more money out of it at the other end.
• The market is pulling the solution from you as demonstrated by referral and engagement

And to be clear, you need to know how the machine works - it’s by design, not by chance.

Small scale? Don’t care. It’s the existence of the value creating machine that matters. The important thing is that it’s repeatedly working.

If it is, you can claim product/market fit.

It’s difficult to do it alone

One of my favourite startup quotes comes from Andy Rachleff, formerly of Benchmark Capital:

“When a great team meets a lousy market, market wins. When a lousy team meets a great market, market wins. When a great team meets a great market, something special happens.“

I like it because it points out that the market is the Oscar winning leading lady of this production, overwhelmingly so (and don’t let media portrayal of successful entrepreneurs convince you otherwise - that’s just availability bias) but team and finance are still non trivial supporting cast members.

Finding the right running mates is going to make your life easier and having sufficient money in the bank to keep you in bread and water will stop you from starving - but remember that neither is true success.

Don’t fool yourself that forming the team or raising money already makes you successful.

Neither of those are product/market fit.

Eight steps

Step 1. Carve out a segment worth fighting for (segment/problem)

Figure out who you want to help and in what broad domain. As great designers oft say: “If you’re designing for everyone, you’re designing for no-one” don't be that person.

Is your segment big enough? No? Then do you have a defendable line of sight to solving adjacent problems for other segments? No? Ditch it.

This is back of the envelope territory. If you spend more than a day on this, you’ve worked too long (ex strategy consultants - I'm looking at you)

Step 2. Listen to the segment (segment/problem)

Get out of the building - your customers aren’t in here. Stuck in a pandemic? Get out of the building virtually. People are often bored as hell and are now masters of Zoom - capitalise on that by setting up video calls.

Don’t assume that your segment has the problem you think they do.

Talk one on one with people in your target segment and for God’s sake don’t drip feed them what you want to hear. If the problem is real and burning (and it better be for you to succeed as a startup) then they will spontaneously raise the issue if you ask them to tell you their domain story.

If they get emotional, start hitting furniture (as one interviewee of a sports subscription did with me) or stop the conversation to call in other people to listen, then congratulations, you’re probably on to something.

How many? Ten minimum. The majority should be spontaneously revealing the problem to you.

Nobody seem quite as interested as you thought? Now’s a good time to pivot - adjust the segment or look for other problem domains.

Step 3. Ideate a broad set of solutions (problem/solution)

It went well. People are spontaneously talking about the same gripe. There are probably many ways to skin the solution cat.

Be good design thinkers and don't settle on one brain child just yet.

Think divergently, think of many solutions. Get others to help. Get your segment to help!

You’re going to need to be able to communicate your ideas some how. Don't build the product, It's too early. Now might be a good time to generate some wireframes or mock-ups though.

Step 4. Second date with your segment (problem/solution)

Get back in touch with the people you spoke to. They’ll be keen because they have a burning problem remember?

Talk to them about your solutions. If they ask to buy the product immediately then you can clock it up as a likely win.

Whatever the result though - don't trust it. Your N is still too low.

Step 5. Build a landing page (problem/solution)

Now you understand the vocabulary and key benefits of your concept (because you were listening during the interviews) - you can properly define your segment/problem/solution sufficiently to validate it.

Build a landing page (see this article on how to do that).

Fire traffic at it from sources where you can target based on your segment and only pay attention to the on page conversion rate (ad click rates are too easy to game).

Conversion rate < 10% after 100 uniques? Ditch it / rework. caveat, yes, having a single conversion rate is ludicrous but this is the general benchmark I use

Step 6. Sell vapour ware (optional product/market)

Reach out to all those people you spoke to and your flush new email list and tell them you’re going to offer the product/service to two lucky denizens.

Get them to write back with their highest bids to be the first. We’re trying to flush out the willingness to pay here.

If nobody writes back, tread carefully.

Step 7. MVP 1 (product/market)

Build a true MVP of your service. Here’s an article to help.

I’m talking about Wizard of Oz style, functional but manual. Don’t write quick bad code (or let someone else do it for you). Build a functionally limited product and do most things manually in the background.

Concierge this. Your two denizens are now royalty. Any time one of them has an issue go Lean manufacturing style and stop the production line, get to the Gemba and keep asking why they had a problem until you solve it for them.

Step 8. Keep iterating and monitor your referral rates

Make sure you have a backlog. Not just for code but for experiments you want to run on the business model.

Don’t run too many at once. Focus on one modification and accept that the result will allow you to accept or reject your hypothesis. Both are success as they move your forwards.

Keep iterating until you find you have a working and understandable little factory/machine that is generating referrals (the purest form of proof I know of).

If your machine is a value multiplier at even a tiny scale then congrats. You've found product market fit. Open some champagne.

The wilds beyond

Whether you now go and raise money to fund your scaling or simply sit back and reap the rewards of your work is of course up to you.

What's for sure is that all the data you collected will make your investor conversations a good deal easier.

You won't need to sell them on a dream, the risk becomes only “will it scale?”.

Oh wait, there’s also a step 0.

Step 0: Define your Zombie milestone

Pick a reasonable time frame (usually 6-12 months) (your Zombie milestone).

If you can't generate your first sticky customers (non churning) in that time frame then chances are it's a no go.

What happens if you ignore your zombie milestone and continue? Well, there's a slim chance that you’ll succeed anyway but a bigger one that you’ll over invest in this concept when there were new fertile fields awaiting you further down the line.

Remember that every minute you invest in a failing concept has high opportunity cost. You could be spending it working on something more likely to succeed.

This is kind of a “Do as I say vs do as I do” step. It’s very hard to drop something and still sleep at night. If you manage to successfully execute step 0 then please let me know, you’re likely a better entrepreneur than me.