Background and context

Opendoor is a real estate technology company focused on making it easy for people to sell their homes and purchase new ones. I worked on the Home Sellers team, where I was focused on improving our offer and closing experiences for home sellers.

We did not have any user researchers when this project started, but we did have a wonderful support team - they were speaking with customers daily, and I was often listening in on those calls. We also had a ton of qualitative user feedback from our NPS surveys that no one had really dived into. So, to start out I decided to dive into our offered NPS to better understand why it was so low.
‍
The primary goals for this work included:

• Improving seller satisfaction with our offer presentation
• Building on a series of improvements to help drive greater offer conversion

Where we started

Diving into the customer feedback, I saw clear patterns almost immediately. We had two huge pieces of feedback: 48% of potential sellers felt our fees were too high, and 36% felt our offers were too low.

Opendoor fees at this time were around 8-10%. The actual cost of selling a home is typically about 10% of the homes’ value. Sellers anchor on realtor fees, which are typically around 6% - but traditional sales also involve lots of other fees like seller concessions, title transfer, notary, and more that Opendoor includes in the service fees.

We also weren’t showing sellers how we calculated our offer, and they had a lot of questions. Realtors typically help homeowners value their home by using data from nearby home sales. Sellers are used to seeing comps, and it’s often how they judge their own homes’ value, even without a realtor.

We had shown comps in the past for just 2 weeks. The previous team saw conversion drop to 2%, freaked out, and pulled the plug. I was skeptical –especially since I saw a lot of comments where sellers were specifically requesting to see comps or let us know about one they think we missed.

What I did

After understanding and aligning on our initial goals and user feedback, I began very early solution exploration with competitive analysis of other sites (Like Beepi, Shift, Carvana) to see how they presented their offers. Taking a handful of different approaches, I put together some mockups to generate conversation among more of the team.
‍
There were a few high level ideas I was exploring:

• Increased visibility of fees and net proceeds: Sellers felt nickled and dimed because the offer showed a nice, big gross offer amount, and much lower down the page we revealed the estimated net proceeds after fees.
  
• Compare Opendoor against a traditional home sale to show sellers what fees they’ll have to pay in either case
• Reinforce our value propositions (No showings, choose your close date, late checkout) to show sellers that even if Opendoor nets them $1,000 less, we might be the better option
  
• Bring back comparable home sales to the offer experience – In the past, the team had experimented with including comps and they watched offer acceptance drop from 4% to 2% and pulled the plug after only 2 weeks. I wanted to revisit this. Sellers were demanding these as they are a staple in traditional home sales.
• “Show our math” Similar homes in your neighborhood sell for around $X - we adjusted your home’s value +$X because you have a pool, but -$X because your lot is smaller than most of the other homes in your neighborhood.
• Including high level market data for your neighborhood - how long are the homes near you on market for before selling? How much over/under asking are they getting?
• Request an updated offer: Other customer feedback included complaints that the offer was too low and one-sided. Sellers wanted a way to share more about their home and request an updated valuation. Our operations team wanted them to include photos.

Feedback and iteration

I hung the initial mock-ups in a meeting room and invited some PMs, Engineers, Designers, and members of our Operations team to give feedback and share insights. I shared the initial design explorations and concepts as well as some of the competitive analysis, wire-frames, and the user feedback we'd been hearing. I gathered great feedback and learned about additional engineering and support limitations our team was previously unaware of.

Some of the highlights from what I learned included:

• Our seller customer experience team was being bombarded with emails from sellers who wanted to talk to them on the phone. They were happy to chat with sellers, but scheduling the phone calls was tedious.
  
• Customer experience was wasting more time than they wanted to emailing users who had no further interest in Opendoor. They wanted better funneling for starters after the initial offer.
• Our existing seller experience to request an offer didn’t ask or provide an option for photos. Our valuations team was working on overhauling the offer request experience, but in the meantime, our appraisers really wanted photos, and when they had to reassess homes, they were even more useful.
  

Final design and results

Our new offer included a bunch of additional features. We started repeating the information we'd gathered during the initial offer request/intake in order to help home sellers to more easily spot any errors or omissions.

After user testing, we also chose to bring back comparable home sales. They are such a real estate staple and sellers wouldn't accept not seeing this information in an offer packet, even if other factors contributed more to our valuation.

The new offer page also shows the Opendoor offer against a traditional homesale to help communicate the costs beyond the fees – in any sale, sellers need to cover repairs, seller concessions, closing costs, and more.

Finally, I created a decline offer option. Previously, sellers could only choose to accept the offer or walk away. Offering an alternative helped us to capture potential customers we'd otherwise lose. This also helped to reduce the hours our operations staff might chase after an unresponsive offer. This gave us folks who were more interested to prioritize for outreach, and it kick started the conversation. (We also finally started capturing photos to help appraisers when re-making offers).

In the end, our NPS rose from -30 to 0, and paired with a 0.5% reduction in fees, the new offer experience led to an increase in offer acceptance from 4% to 10%