Seven lessons learned failing my first startup
In September, I closed down Wine List. We were trying to do for wine what Duolingo did for language, but we failed.
My background was in growth for DTC startups. I’d seen what acquiring 10,000s of customers looked like. I’d seen what failed experiments, cultures, and businesses looked like too. I should have been on good footing.
So what went wrong, and how would I do it differently again?
1. I made too many excuses for PMF
We never found product-market fit — I knew this, but I ignored it.
If there is one hill I will die on when talking to other early-stage founders it’s this one: stop making excuses for not having product-market fit.
Sean Ellis came up with a framework for PMF: ask your customers how disappointed they’d be if you disappeared. If 40% tell you they’d be very disappointed, the framework says you’ve got it. If it’s less than 40%, you don’t.
We ran this process with our entire customer base four times, and with a dozen cohorts on a monthly basis. We never got above 25%.
But I said, ‘we’re acquiring customers at X, once we hit scale, that would mean they are profitable after three months.’ (Excuse)
But I said, ‘our customers are typically older/British, how many of them would ever describe being ‘Very Disappointed’ that a company disappeared’? (Excuse)
But I said, if you look at this one cohort — we do have PMF. (Excuse)
I spent two years making excuses for not having PMF. But I knew deep down we didn’t have it.
How did I know?
2. I was terrified of turning off marketing
Every month I’d write an investor report and there would be some good stat in there. Either revenue or active subscribers were almost always going up. But I was terrified of turning off marketing.
I knew that if I did, churn would be around 20%. Revenue would drop in line.
I was terrified that even though our retention curve was flattening, it wasn’t flat. I was terrified because the margins never seemed to quite line up in practice as they did on paper.
Marketing hid many of these problems, and I was terrified every day of turning it off.
3. I acted like we had a future state of LTV/payback rather than today’s
Wine is an economies of scale business — and at low scale it has sucky margins. First six months, we had negative margins, next year was 10–15%, eventually we got to 35%.
“When we hit 5,000 customers,” I’d tell investors, “we will be able to push those margins to around 60%. And based on that, payback will be three months, and we’ll be at 3:1 LTV:CAC before the end of the year.”
Future states are important. It shows future viability of what scale means.
But I let myself think too much in the future state, and not enough in today’s.
In addition, lockdown meant 75% cheaper CPAs and about 30% better retention. Compound that with the future margin mindset, and today’s business burnt through cash way too quickly.
4. I fell in love with my first startup idea
I always thought of myself as an entrepreneur waiting for an idea. That’s a bad way to come up with startup ideas.
The process was: go through day-to-day life, wait for an idea to hit me, create a solution, experiment and validate, raise money, go.
“Ideas are worthless, execution is everything”
I’d often say this for years — and I retreat on it now. Ideas are important.
David Perell analysed the Simpsons writing process, and they said they rushed to the end of the first draft because they knew it wouldn’t be good. You need to flush out the crap, before you can get to the good stuff. Create an imperfect world, then improve it.
With Wine List, I found a problem and didn’t bother flushing out. I created the imperfect world, and ran with it from there.
Whatever business I start next, I’ll start with dozens of ideas, create success criteria up front, and test those ideas against them.
5. I fell in love with a solution — not a problem
“When I studied wine, my enjoyment of it went up by an order of magnitude,” began each pitch, “but the way I learnt was broken.”
I fell in love with the joy I felt having experienced a solution.
We always struggled to really communicate our problem-solution to our customers in an exciting way.
We found ourselves doing two jobs:
(1) teaching people why they should learn about wine, and
(2) why we would be the best way to learn
Matt Lerner told me years ago, he worries when he hears about startups ‘educating the market’.
In marketing, we know that attention is nanoseconds long. Grabbing it is difficult. Grabbing it and then telling that person they need to adapt their mental model to buy from you? That’s going to be almost impossible.
You see this in different ways: like shitty click-through rates and conversion rates below 1%.
And if you asked 100 wine drinkers ‘do you wish you knew more about wine?’ You’d find almost all of them would say yes. But if you asked if they would shift something in their life to learn, the number would go down a lot.
Sure, people wished they knew more about wine — but we never validated that problem. We validated a solution.
6. I lost my risk tolerance as soon as it became my money
In my life as a growth marketer, I almost always pushed for more budget and more risk.
“How much do you want to spend?” one boss once asked me, “how much do you want to grow?” I responded.
I had a very high risk tolerance.
But when it was my money, and I had to personally answer to my investors, that changed.
Being risk-averse in a startup is dangerous. It stops you making bigger bets like product changes or hiring new people.
I ended up optimising for runway rather than learning.
7. I lied to myself about what was needed to launch a new product
In September 2020, we tested a prototype of a 125ml-sample based ‘wine through the post’ product. It worked great. But we didn’t sell our first box of this until April 2021, why?
I told myself three lies around this product:
1. We need our own warehouse
2. We need to finish fundraising
3. We need to have a full time ops lead in place
We finished fundraising, and put an offer on a warehouse. We started hiring for the person to lead it, we met candidates, went back to the spec, and refined it.
Before we knew it, it was March. Six months after our first prototype and we still hadn’t built this product — or tested it with the market.
I took a holiday and realised that I didn’t need any of this. Getting this live was a series of tasks. I got back, and outsourced product design, logistics, and first build. Three weeks later, we had a product in market.
In May, we acquired over 1,000 customers to By The Glass at a £10 Facebook CPA. Our landing page for this converted at 6%, vs 0.8% for the core product.
I told myself three lies about launching this. I thought we needed conditions to be perfect.
Had we outsourced earlier, we could have had our first products out before Christmas 2020, and our future could have looked a lot different.
Bonus takeaways
Here’s a few final, smaller things which are worth pointing out.
i. Build recreational / celebrating time in with the team more
As a manager, I was great at building recreational time in with the team. As a founder, this slipped by the wayside.
And when I look back now, it’s not the day-to-day stuff that jumps out but the memories of our team together. The Christmas party, the time we went go-karting, the day we went around vineyards, and the warehouse party.
These things matter. They’re fun. The bonding that goes on with them helps cement relationships between colleagues that means when things are tough, they’re that little bit easier.
I wish I did these more, I wish I did them every month. Next time, I’m going to celebrate with the team more.
ii. No-code is great for MVPs, not so good for full products
“Product hacker is the new VP Engineering” was a post I was once going to write. I am glad I didn’t.
No-code is amazing. It means non-technical founders can build entire first products. But, we found as we were trying to build v2 of our site + product, it took just as long with no-code as not, and had far less capabilities.
iii. Continue to over-index on founder-investors
I had over 20 investors in our pre-seed. Half were founders or former founders. When Wine List closed down, one of them — Basubu founder, George Burgess, — called me up to say “well done, you’ve done it now, you’ve had a failed company, and you know it doesn’t matter.”
That was one of the best things I heard in those early days after announcing closure. He’s right. The fear of failure was always a hindrance — to what degree I don’t know, but fear of failure made me make worse decisions in many ways.
I optimised to ‘keep it going’ rather than make faster decisions that would have bigger impacts.
George was right, I’ve done it now, and I’m liberated from that fear again.
Those sorts of insights and feedback felt unique from the founder-investors. Other investors had value in different ways for sure, but the founder-investors were the ones I felt I could most easily turn to for help.
Thanks
I’d like to thank everyone for joining me on that journey. Thank you to the team old and new. Thank you to all of the investors who supported me every step along the way. Thank you to the customers and the community for joining us on it. And thank you to friends and family for putting up with having a founder as a partner, and friend.
For this article in particular, thank you to Matt H. Lerner, George Burgess, and Peter Sigrist for revisions, suggestions, and feedback.
Want to stay in touch?
Some final thoughts — edit
Sadly as part of this process, this business ended insolvent and as a result we owed a handful of producers and suppliers money. It’s the real downside to an end like this, and the part of the process that I personally feel most sad towards. If you liked any of the wines we sold, please go and support them by buying their wines wherever you can.
