As the first Product Manager at Knock, I was responsible for the development and release of two mobile apps and a set of internal tools. During that time, I helped lead the expansion of the product and engineering team from 6 to 60 employees and oversaw product strategy for the entire customer experience. Knock saw significant growth, reaching a valuation of $2 billion before it fell victim to the latest real estate challenges. You can read more about that here…
Here are a few problems I had to solve…
Is that house really worth that much?
Knock allowed people to “swap” their house for a new house. This meant that they could buy a new home before having to sell their existing home and they could do it all with cash offers. To do this we needed to create customer onboarding tools that integrated with our machine learning models to help mitigate our risk and scale our business. As good as our models were at understanding the value of a home we still needed a trained individual to physically inspect the property. The challenge we faced was how to efficiently scale this process across numerous markets while maintaining a personal touch.
What we knew…
Potential customers expected to get an immediate budget number for a Home Swap (this allowed them to start shopping for a new home). If they didn't get a final number from us in 48 hours we would lose them as a customer.
Competitors were catching on and starting to copy our model so we needed to build a moat.
Whatever we did to solve this needed to scale in new markets.
We had a lot of inquiries from local real estate agents that were interested in what we were doing.
What we did…
We created a mobile app for local real estate agents that allowed them to accept and conduct these home visits in an uber-like way. Prospective customers could self-schedule a home visit at their convenience. Local agents could download our app and conduct home visits (gigs) gathering data in real-time that our teams could use to confidently finalize deals with our customers.
We were able to release an MVP to one of our markets in a little over a month. As expected we saw a quick increase in top-of-the-funnel conversion. Customers were choosing to move forward with our service closer to 70% of the time instead of just 50% and we were usually able to have someone visit the property within 24-48 hours. However, there were still questions about how we could scale this since agents needed to be trained and there were large variations in the quality and accuracy of the reports. Over the next quarter, we made many improvements including how data was captured (images and labels), leveraged GPS to decrease the number of late/missed appointments, improved usability to make it easier and quicker to complete the home visit, and implemented a feedback mechanism that allowed us to prioritize visits to agents that created the best reports.
How did we do?
In less than 6 months we rolled this app out to all 8 of our major markets and rolling out new markets required minimal engineering resources.
We had a growing network of agents that were engaged with us and supplementing their income by doing home visits. This significantly increased our referral business.
We were able to get very accurate home valuations back to customers in less than 48 hours and in some cases the same day. We also drastically reduced major valuation errors where properties were in horrible condition or had odd functional obsolescence.
We gathered millions of labeled images that we could use to improve our ML models. This was data that nobody else had.
Of course, it wasn’t all perfect. Freeing up the top of the funnel and smoothing the onboarding experience soon led to issues downstream.
2. Indigestion
I once heard that companies don’t starve to death they die of indigestion. To this point, most of the work I had done was focused on growth or expansion, and the thought of having too much demand was a welcome worry.
After various growth initiatives, things were looking good and revenue was growing at 53% QoQ. The top of the funnel was strong but that growth started to cause problems downstream. The Knock brand had a spotless reputation among various online rating websites but that was starting to tarnish. It seemed like every day a new customer was sharing a story about our service that was not what we envisioned or had come accustomed to hearing. I was asked to restructure the teams and focus entirely on improving the customer experience. Growth had always been our focus but our success had created new problems. I asked my team to meet with partners, customers, and anyone else willing to talk. We formed a research arm within the design organization helping us formalize and centralize the feedback.
What we knew…
We had a lot of happy customers but also a growing segment of very unhappy customers. There wasn't much in between.
Unhappy customers were feeling confused and abandoned.
There was a lot of disagreement on the actual problems and also how to solve them.
Our service and support team was overwhelmed due to the complexity of the transaction.
Our partner agents had some skin in the game and wanted to help but didn’t know how.
What we did…
We quickly implemented in-product surveys and various other feedback channels that allowed us to continuously discover opportunities and issues. We implemented NPS at key points in the process which helped us better understand where to focus efforts. NPS became our north star metric and was reported and shared weekly. It’s not perfect but it was a single measure with an industry benchmark, and it often came with additional comments that were very insightful.
Below is an example of some of the data gathered and analysis conducted. This allowed the product teams to determine where to focus their efforts.
NPS (touchpoint and brand)
Correlation of other customer data points to reported NPS score
With additional insight, each team went to work on ways to improve their products. The project focused on improving communication and streamlining processes in order to deliver a higher quality customer experience consistently. This was achieved through better deals assignment and automated introductions to the appropriate team members. Partners were also included in the communication through a mobile app, enabling them to address and resolve any issues quickly.
How we did…
As a result of these efforts, we improved the Net Promoter Score (NPS) by 60 points within 5 months and enhanced the brand's reputation. This put us in line with the industry standard for companies operating in the mortgage origination space. Complaints regarding “communication” dropped significantly as customers began reporting that they felt more heard and that Knock had their back. Additionally, our operations teams were more aligned and reported happier work-life balance while increasing their case load.