DoorDash: The Power of Customer Cohorts

Written by
Ross Andrewsarrow icon

DoorDash: The Power of Customer Cohorts

Written by
Ross Andrews

DoorDash filed what was one of the most hotly anticipated S-1’s last week (you can see it in full here) setting a clock on their IPO. DoorDash along with other similar delivery services such as Uber Eats and Postmates, have received a lot of scrutiny of their business model and economics over the last few years. But how refreshing to see a detailed cohort breakdown of DoorDash’s customer base the evolution of their unit economics over time, let’s dive into it and see the power of customer segmentation and cohort analysis.

An interesting graph which is feature prominently in the S-1 is a breakdown of Contribution Profit by customer cohort which you can see here:

contribution profit by cohort

First off, quick house keeping. DoorDash defines Contribution Profit as gross profit minus sales and marketing costs, Doordash’s version of retention profit or their cut that is left over after they have spent on keeping customers. In the above graph we can see that after incurring a loss from their customers in year 1 in the 5-10% range, their customers actually produce a contribution profit in years 2, 3, 4. The big question, what is driving this? It’s an important indicator of long term success. For DoorDah it’s two things.

https___bucketeer-e05bbc84-baa3-437e-9518-adb32be77984.s3.amazonaws.com_public_images_c80a9ee0-78a1-4663-a189-48e44600e0d7_637x353.png

We talked about the contribution profit being gross revenue minus sales and marketing. From the above graph we can see that after year 1, the % of sales and marketing expenses that DoorDash spends on those customer cohorts drops 50-70%. The decrease in expense improves the unit economics of each customer to the point that in years 2 and beyond as we can see in the first chart, they are now profitable operating within the core business model, that is taking other operating expenses off the table. The second factor, increased order volume over time.

gov-by-cohort

This chart shows GOV or gross order volume grows year over year from the initial cohort year, this is DoorDash’s way of defining expansion revenue which is a key driver of solid unit economics in SaaS. Given that DoorDash earns fee’s on each transaction in their app, so as gross order volume goes up, so does their fee’s and revenue. If sales and marketing expenses drop in years 2+ and gross order volume increases this create a clear path to positive unit economics for DoorDash.

Now there are two unknowns here. DoorDash still operates at a loss, even with the cohort breakdown painting a path to positive unit economics, will other OpEX ever come into line that they can operate at a profit? TBD. But more importantly over the same period, what does customer churn look like? What does revenue retention look like for those customers that don’t increase their GOV, or is DoorDash masking the loss of GOV through customer churn by pushing expansion revenue from existing customer creating a false bottom. Eventually you will run out of ways to increase existing customer revenue and with high customer churn, you will hit a wall in terms of revenue growth and those positive contribution margins. Now that isn’t all doom and gloom for DoorDash, but if it is the case, it could limit their total potential market size which directly affects revenue and price of their shares.

Overall I have to applaud the DoorDash team on their S-1 👏. The cohort analysis is a great educational tool for founders and new startups of a financial model to build on, using sales and marketing expenses to drive customer acquisition at a loss and then through decreased expenses and expansion of revenue to have a healthy CAC payback and healthy profits from the customers life-time value. Too often recurring revenue or SaaS startups don’t take the time to look at segment or cohort  level data which often times can uncover streams of strong unit economics and uncover strong growth opportunities.

Sign Up For Our Weekly Newsletter