Discover more from Early Stage Growth
Choose Your Own Adventure: the Big Bet we played at Wine List
Faced with no product-market fit, we had a couple of major plays during our time. This post details the background and decision-making behind one of those.
When I talk with early-stage, pre-PMF founders, I often talk about 'Big Bets'. Determining the order of magnitude for a Big Bet is a tricky task.
At Wine List, we introduced a "By The Glass" product iteration. This was, operationally and strategically, a substantially big bet.
Ultimately, it didn't save us. But I want to re-examine the decision to do that Big Bet today.
This is the first post of this kind, and I may well do more of these if they prove to be interesting. For this post, I've used previous Investor Updates, historic notes, old data I've still got access to, and my memory to put it together. I've tried to be as accurate as possible, but note that this was written with hindsight and I am sure my memory has been selective in place.
Also if you wanted to view this as a Choose Your Own Adventure story, then I’ve signposted the point to stop reading so you can consider how you might have acted at that time. And please do share – hit me up direct: email@example.com.
Problem: rising CPAs, moderate retention, and low early margins
First then, the problem. Let's take a look at this backdrop.
Rough timeline notes:
March: went full-time. First real month on the business. Light FB testing. Then lockdown happened. Was brilliant for business.
April: pushed FB spend. FB CPA went from £18 to £53. But we had so much PR/organic/affiliate/partnership that blend was steady at £24
May: push FB again, managed to have another strong month there. Couldn’t increase non-paid at the rate needed to keep blend down.
June: FB CPA went up again, pulled back spend. And again non-FB couldn’t keep blend down
July: pull back on spend a lot as worried about that blended CPA increase. FB CPA stays same.
September: costs massively spike. I noted at the time this felt very platform-driven.
October: pulled spend again.
So as we all know there is no universal ‘good’ CPA. It all comes down to your unit economics. So let’s first look at what retention was.
We were averaging 75% monthly retention at this stage. Now given by mid-summer lockdowns were easing, it was difficult to forecast. After all our 3 month customers were now stickier because of longevity. But the newer cohorts would presumably have different behaviour.
Still ROAS-wise, FB was only ROAS1+ in March & May.
The real question of course lies in payback and margin.
The margins of a wine business
We were buying from existing importers up until mid-summer. This is the de facto way that almost every wine shop/retailer operates. On a bottle, there’s about 30% margin usually.
Remember, however, that our proposition was focused on learning about wine. Every box contained learning materials. The box design was reflective of our £39 price point. And so by the time you had stored the wine, packed it, printed off booklets, and added in materials, shipped it, and taken into account breakage (wine is high), your margin looks a little different.
Now, we had a particularly bad 3PL which reduced our margin more than it could have done, but even without that, you were looking at ~5-10% margin by the time everything had gone out the door.
As someone once said, “If you want to make a small fortune in wine, start with a big one.”
Let’s look again at out unit economics with an 8% margin.
We always knew we would want to become a wine importer at some point (which improves margin), but this became clear it was a necessity.
But in truth there were a few problems here rolled into one.
How we drew up hypotheses, experiments, and Big Bets to deal with the problem
Operational action plan
There were two operational drivers to get that 8% figure up.
First: we would have to become a wine importer. Importing wine meant that instead of buying wine from existing distributors, you could go to the source and drink up both importer + retailer margin. This is what every massive wine company does from Naked to Virgin to Tesco.
If we became an importer, that bottle margin would go up from 30% closer to 55-60%, which is fairly healthy.
Second: we would have to own distribution. The aforementioned 3PL were killing our retention with the volume of errors they were producing. We couldn’t be the type of company we wanted (one which if you ordered at 4pm, you’d get wine the next day). And the cost of fulfilling yourself was much much lower than using a 3PL.
Those two operational changes together would improve the shipped margin closer to 35-40%. 45% in some months when things were firing on all cylinders.
What does that do to our payback?
Well for a £30 CPA now pays back after three months, and even some of the punchier CPAs we were concerned about paid back after six months.
Hiding in that retention figure
75% month over month retention initially felt like a strong starting place.
Except that means by month 3 and 56% and month 6 you’re at 24%. These were both way lower than Lenny’s ‘good and great’ retention pieces1.
So why was retention bad? That’s a lagging indicator of product-market fit.
So we started doing PMF research. The research at a numeric level reinforced the data we were seeing play out. We didn’t have PMF. If we were to close tomorrow, the vast majority of customers wouldn’t care.
What the PMF research revealed
We followed the Superhuman template2 and discovered a few things.
Our high expectation customer:
Wanted to taste new varietals, new wines, and learn by tasting
Our anti customer:
Wanted to find good cheap wine and just wanted ‘good’ wine given to them
Of those customers who did love us, the things they loved were:
75% listed discovery of new wines/varietals/grapes
25% listed education
5% listed social activity
From there, we did more customer interviews. Really dug in to the discovery + the education pieces to understand how we could improve them.
We narrowed down the action list in the following areas:
Video guides, video tastings, rewatchable video content
Quizzes, more tasting cards, tasting paper, a filofax to keep taste cards.
Ease to buy more wine (most frequent response), more of the same bottle, more similar bottles, more X wine, more Y wine. More wine was a frequent response.
Events, private tastings, dinners
Now, if you wanted to play a Choose Your Own Adventure game, then now would be a good time to pause.
Sit down with a pen and paper, and draw out where you would take this based on the feedback above.
The three small bets we made to improve our position
Eagle-eyed amongst you may have read those above tables and thought ‘but wait, what if you pushed that AOV? If you could double that, surely you’d be in a great position.
Indeed, even a £60 AOV – likely just one extra bottle of wine sold per person – would turn the unit economics around too. As an importer, that would mean a 3-4 month payback for £50 CPAs, with healthy LTV growing beyond there.
And of the people who loved us, 75% listed discovery. We thought we were in the business of education, but in fact we were predominantly in the position of discovery.
Discovery small bets:
Start selling ‘surplus wines’
Sell additional wines that aren’t the core subscription wines
Introduce case option of 6 bottles
All three of these while also helping fulfil on our PMF issue, would also help fulfil our unit economics issue.
Thinking about Big Bets
A lot of the above would work fine if you were building a lifestyle business. Likewise if you were a retail op that wanted an extra revenue stream. If you could acquire 750 customers, selling £50-60 of wine a month at 35% margin, you’ve got £14k of profit each month. Slow growth, organic acquisition, service those customers well and they’ll be sticky.
But I wasn’t building a lifestyle business. I was determined and convinced that there was an opportunity which could reach 1 in 5 of the UK’s wine drinkers, by helping them double down on knowledge. Or I was trying to anyway.
And so, I needed a Big Bet that would let us land in orders of magnitude far greater than where we were. Our ultimate goal was still 200x revenue away.
We had a few routes available to us: one would be to go ‘full app’ – minimise cost of acquisition to <£5, deliver wine content to them maybe as a subscription, and upsell them wine. This was the route we eventually tried raising money for a year later.
But in the short term, we weren’t yet convinced it was the best route. I went back instead to my very first idea for Wine List: a box of taster glasses. 5-10 small/half glasses of wine, rather than bottles.
It would over-service the ‘discovery’ people. And would act as a vehicle for even greater learning too. It felt like win win.
That would mean building a way to rebottle wine into smaller containers while protecting them against air damage. That is scientifically very hard to do. It is logistically and operationally even harder.
By The Glass ended up taking up 40% of the entire company time to get it to market. That was a substantial endeavour. We prototyped it, sure. We gauged interest. We tested the price point. And then we did a slightly larger roll out to kick the tyres further. But we wouldn’t fully know until we had it in market. The opportunity cost of By The Glass was substantial.
When we eventually did release it to market, we got Facebook CPAs at about £10 on a ‘free glass, just pay postage’ offer as a first box on a subscription. At the time we couldn’t get the core Facebook CPAs below £80-100. That seemed to prove out the ‘market’ side of the equation. But we had significant operational issues that I think contributed in large part to the massive churn we had.
By The Glass was indeed a Big Bet – and a much larger one than I initially imagined. I’m not sure how I’d have aptly understood how much operational cost would be involved in getting it to market.
With a fresh and outside pair of eyes on this, I’d love to hear from you to hear how you would have approached it differently.
And if you are stuck with no product market fit, and want to discuss some Big Bets you can be making, I’d love to discuss these with you. My 15 min call slots are always free and open (and we can chat more after).
“How Superhuman built an engine to find product market fit” First Round Review