How a single adjustment increased Opendoor’s revenue by $1B in 2021
Identifying a gap in our re-engagement process that led to a significant increase in customers selling their homes to Opendoor.
Context
In my previous post, I discussed how Opendoor used Multiple Listing Service (MLS) data to accurately track home sales. We also talked about our north-star metric called "true seller coverage," that we defined as:
Internally, we categorized sellers into two buckets:
New sellers who sold within 30 days of receiving their initial offer, and
Re-engaged sellers who had received their initial offer more than 30 days ago.
From a total volume perspective, each group typically contributed about ~50% of total homes that were sold to us in any given month.
Goal
In early 2021, our re-engagement team consisted of Robin (engineer), Tyler (data scientist), Joey (lifecycle marketer), and me. Our objective was to increase the number of re-engaged sellers who sold their homes to Opendoor. We decided to use the true seller coverage metric, which measures the percentage of sellers we successfully re-engaged compared to total sellers in the market.
Re-engagement process
To reconnect with customers who hadn't yet sold their homes to us, here was our rough documented process:
We emailed them every 7-14 days during the first 90 days and then every 14 days thereafter. Most of the content was related to the process of selling & market trends.
We generated a new offer every 28 days, typically towards the end of the month.
We also sent them direct mail based on a model that predicted if they were looking to sell their home.
Discovery
To analyze the data, Tyler evaluated seller trends using a variety of filters, such as geographical location, the timing of their last transactions, and when they initially approached us, among other factors. During that investigation, we found that about 90% of the homes purchased from re-engaged sellers came from those who had received their first offer within the first two years. There was a steep drop-off after the two-year mark (see sample data below). By looking deeper at the activity of these sellers who were beyond the two-year mark, we realized we were sending them no email communication at all.
We traced this back to a decision that was made to stop sending new offers and emails to customers after two years. The decision was made sometime in 2020 as preparing offers by valuation experts manually (which was no longer the case) was time-consuming, COVID had restricted operations and seeing higher-than-average unsubscribe rates for that group.
Solution
In 2021, since we had automated offers for most sellers, we tested removing the two-year filter and sending emails and offers to all re-engaged customers. We tracked conversion as the primary metric and unsubscribe rates as our secondary metric. Our team was optimistic about the experiment but expected only a modest initial impact.
Impact
The test, which lasted for a month, exceeded our expectations. In that period, we doubled the number of contracts from the re-engagement pool, leading to an annualized increase of $1B in home purchases (which we sold to get >$1B in revenue). The impact persisted for several months after the test concluded, and we eventually rolled out the change to all sellers with only a slight increase in unsubscribe rates.
The trade-off was entirely worth it given the size of the impact. Our primary challenge was handling the surge in customer sales, as our sales & support team was unprepared for the sudden influx.
Key Learnings
Analyze your data to discover hidden opportunities for growth.
Be open to revisiting and adapting previous decisions – especially when fundamental limitations have changed.
Test real-life data compared to arbitrarily hard-coded numbers in your logic.
Prepare for growth by ensuring all teams can handle increased workload.
Continuously refine your re-engagement strategy to match evolving customer habits.
Conclusion
Our experience at Opendoor demonstrates the power of using data to drive growth. By identifying and addressing gaps in our process, we doubled our contracts with re-engaged customers and significantly expanded our customer base. We hope this inspires you to dive into your own data and make impactful changes that will help drive success for your business as well. If you have any examples of simple adjustments in your own business that had tremendous impact, please feel free to share them in the comments below.
Thanks to Alexey, Shimolee, David, Sarah, and others for providing feedback!