We’re often in awe of big companies that appear to be overnight successes. We think, if only we’d had that idea then we too could have become billionaires. But the reality is for every AirBnB or Uber, there were a slew of similar businesses that never made it. I remember, in my agency, we had a client who contracted us to build them a social network where friends could connect online and share updates and photos. This was in 2005.
I don’t know about you, but I created my Facebook account in February 2007 at the same time as all of my friends seemed to be doing the same.
Our client who had the same idea as Zuckerberg? Who knows what happened to them. The URL is now owned by someone else.
As we all know, having an idea, even an original one, isn’t worth much in itself. The success comes from the execution and when you look harder, you’ll see that it’s unlikely anything was ever an overnight success.
Even the concept of an ‘overnight success’ suggests that it was down to one lucky ‘thing’ that catapulted them to stardom. Often I have people who want to have a call with me and I can tell straight away that they are hoping that I can give them the secret, the ‘silver bullet’ that propelled ScreenCloud to the successful business it became from our origins as agency owners. But it would be like saying that a marathon runner could trace their success to a specific training run they did 6 months earlier when in reality it was the culmination of lots of (probably unremarkable) training sessions that contributed to their ability to win a race.
Small Bets
Did you know that AirBnB launched their site three times before they saw growth? They were so broke that the founders were living off breakfast cereal until they could raise more money. Their eventual ‘overnight success’ came from a hack where they were able to automate the re-listing of everything on AirBnB into Craigslist. But this came as a result of multiple experimentation and sophisticated programming.
By the time that they saw the fruits of their labor paying off, they had also discovered the value of running multiple small bets. These are micro experiments that seek to move the needle, but perhaps by only a small amount. Today, AirBnB run about 2,000 experiments a month.
Let’s say that you have a current conversion rate of 10% of leads to paid customers. And each of your customers pay you $1,000. In a year you get 10,000 leads. You will get 1,000 customers and $1m in revenue.
If you can increase the conversion rate by 5% to 15%, you will get 1,500 customers and $1.5m in revenue.
But conversion rate is one element out of many that you could improve over time. What about if you could increase the leads at the top of the funnel by a few percent? What about if you rose your price by a few percent? What about if you reduced churn by a few percent? What about if you increased the average order value by adding new services or encouraging people to add extra licences?
Within each of these elements are more sub-categories of growth potential: converting leads from 10% to 15% might seem like quite a hike, but not if you look at making tiny improvements to each stage of the funnel as they compound. The higher up the funnel the improvement, the more it will impact the eventual outcome.
Creating a Growth Squad
AirBnB’s 2,000 experiments a month is a lot even for established product companies. Part of the problem is the manpower required to implement them and part is then being able to track what is working and what isn’t. It also helps if you have enough data to give you meaningful results which you’re more likely to do if you have a B2C offering. The other issue is making sure that multiple experiments don’t overlap. If for example, you did a price rise on a page where you also changed the call to action, how would you know what was the thing that caused the change?
Does this mean that for B2B SaaS businesses you shouldn’t bother? Not at all. Most startup SaaS businesses should be able to comfortably run 4 or 5 experiments a month. And this can be enough to deliver real results.
The important thing is to include people from across the business. AirBnB’s Craiglist hack wouldn’t have worked if they hadn’t included developers who were able to automate the reposting of AirBnB listings. If your aim is to increase conversion rates, that could come from improved marketing efforts but it might also come from better onboarding or different messaging within the product itself.
When I was managing growth at ScreenCloud I created a Growth Squad. It consisted of people from sales, marketing, customer success and product. We met once a week and agreed what experiments we were going to run and who would be responsible for running them. Looking back, forming a Growth Squad was an experiment in itself and running it was a bit of trial and error, but the idea was sound.
I made a small price increase where we went from $20/screen/month for monthly customers with a discount to $16/month for annual customers, to $20/screen/month for annual customers with a premium of $24/month for monthly ones. Effectively it meant that we could keep our headline ‘$20/screen/month’ price, but it became the minimum rather than maximum price. It had zero impact on conversion, but it did add about $650k in ARR over a 7 month period (if my memory serves me correctly) that we wouldn’t have had without the rise.
How do you make small bets?
This is a massive subject on its own, but the way we did it was broadly like this (from memory a lot of this structure came out of the book Hacking Growth by Sean Ellis):
1. We created a Google Form that anyone in the company could use to submit a bet. We asked them to rank each bet out of 10 by:
Likely Impact
Confidence that they had it would work
Effort (how much time and money would it cost us to implement)
That form then populated a spreadsheet with all the rankings included.
As a Growth Squad, we took those ideas and discussed them (along with our own) and agreed a priority list based on the average of the ranking scores. If something wasn’t probably going to make much of a difference and would be a massive time sink to do, then it would take second place to something that was low hanging fruit that could make an overnight impact.
We grouped ideas around hypotheses. An example might be that we had the hypothesis that we could increase conversion through the funnel. Every idea that was related to that was grouped together in Notion.
We then defined the experiment: what we would do, what we were looking to prove, what we would measure and over what time period.
We then ran the experiments and reported back to the group each week with any updates.
One thing about small bets is that most of them won’t have any impact, or they will have a small but statistically insignificant impact. But to my mind, even small improvements add up over time. What I wasn’t able to do, but wish I had, was to take a view that, say, over 6 months, the overall impact of all of our experimentation was X% or $Y. A bit like the marathon runner being able to look back and see that their speed had improved by a few seconds per mile over six months of training, even if they couldn’t say exactly whether that was down to speed training or increasing their overall mileage or, most likely, a combination of the two.
I believe that small bets can be a really powerful growth tool in your armoury but it takes time, commitment and patience to make it happen.