Retaining Customers in a Marketplace
Welcome to Samaipata Ventures’ blog, a VC fund specializing in market-place and disruptive e-commerce models in Southern Europe. If you don’t want to miss out on any quality information, follow us here on Medium, Twitter or suscribe to our newsletter.
Everyone talks about retention. To get it right is tough; to do it really well raises you to an elite of truly famous companies — the Snapchats and Tinders of social networks; the GrubHubs and Wallapops of marketplaces. People also talk about the failures of companies that push for growth instead of working on retention. Many of their users fail to return to the platform the month after they first log on. Often these companies acquire low-value customers through vouchers and huge discounts — customers they never see again once they switch them to the regular price. This money could be better spent on product optimisation and establishing the right sources of customer acquisition.
While user growth is important to demonstrate product-market fit and scalability, in many cases a more important measure of a business’ value is retention. And retention can sometimes be a weakness for young, fast-growing startups. It shouldn’t be. Now that high cash burn is out of fashion with many VCs, there is greater focus on unit economics, operational profitability and cohort performance. User cohorts (sometimes also supply side cohorts, or even rider cohorts, in the case of a three-sided marketplace like Jinn) apply especially to marketplaces because of their structure — typically a marketplace works better when there are more people on either side of the equation. If you log on to eBay and there is only one seller, it is unlikely that you will visit the site again. If there are more sellers each time you log on, with more variety of product, you are likely to come back sooner.
The true signal that a marketplace is delivering real value to its users is when its cohorts start to trend upwards again — this is the inflexion point where the network effect takes hold and users become increasingly engaged, returning more frequently to the platform. For example, with on-demand ride hailing services, the marketplace becomes truly useful when there is enough liquidity of supply for a car to only ever be a few minutes away. For any buyer in a marketplace, high liquidity can mean consistency of service quality (e.g short wait times) or variety of product (e.g. choice between different price or quality standards). For the seller, high liquidity increases demand and therefore volume of sales. Often the inception of the network effect goes hand in hand with an increase in organic customer acquisition and decreasing Customer Acquisition Cost (CAC), as word of mouth fuels growth.
At Samaipata we analyse retention as a function of two things: a) % of First Time User retention, that we call P. And b) number of buys per returning FTU, that we call Q. P is a commonly accepted measure of user retention — a percentage of how many users come back after one, two, three months/weeks. Q is a measure of how valuable these users are to the marketplace — a great marketplace will have Q that increases as the platform arrives at maturity and users become more engaged or change their purchasing behaviour. P x Q is average orders over time / returning FTUs. This is a fundamental measure of how effective your marketplace is, and should be flat once retention flattens out. If it is increasing, great. If it is declining, your marketplace is in trouble.
The relationship between P and Q is often a function of the industry as well as the product (e.g. if people eat once a year, you have no retention). For some industries, food tech for example, P may stabilise and Q may increase as remaining users change their eating behaviour and start ordering more delivery meals. The Holy Grail is for both P and Q to increase as engagement and purchase behaviour both change positively increase revenue, but this is very difficult to achieve as it is unusual for P to increase.
So what drives poor retention? As we mentioned above, marketplaces that push for growth using the wrong incentives, such as vouchers or heavy discounts, are bound to discover a cohort cliff when the low-value customers churn off the platform once they have used the initial discount. Marketplaces with poor operations or poor product experience have similar retention problems — customers churn after they have a poor experience, often without a word of complaint to the company (complaints only represent the very worst of weak user experiences). Unless a company keeps a watchful eye on its Net Promoter Score, patchy operations and low customer satisfaction can go unnoticed and kill cohorts. You can spend money on acquisition but you’ll be pouring water into a bucket with a hole in it.
There is one specific reason for low retention rates that we think is worth special attention — the marketplace leakage problem. Stay tuned for more on leakage in the next instalment..