Redesigning Abodewell’s Offer Request Flow

Abodewell is the simplest way to sell your home. We recently set out to redesign the way people request offers on their homes to improve customer experience, increase conversions, and solve internal pain points.

Overview

The Offer Request Flow (ORF) is where homeowners can request offers on their homes on the Abodewell website by answering questions about their homes and sale timeline. Our Data Science team then generates valuations on homes, and sends offers to homeowners in under 24 hours. Our existing ORF is incredibly simple and easy for homeowners to use, but as we were growing, we began noticing many limitations. So we set out to improve the process for home sellers to request and receive offers, in hopes of increasing conversion from visitor → requestor, and requestor → contract signer.

PROCESS

Started by doing competitor audits, and studying the UX of other form-heavy products like Turbotax and Policygenius. Studied customer email correspondence, reviewed analytics and session recordings to better understand how users are interacting with the flow. Design also worked with our Customer Experience and Data Science teams to better understand the problems in our internal workflow, and how we could address some of those issues with the redesign.

Platform

Web app

Timeline

April-May 2019. Engineering build in July 2019.

role

UI/UX

Team

Sophie (UI/UX), Natasha Alia (Visual Design). Milan Kodali (Product), Marci
McClenon (Engineering)

PROBLEM – Customer Experience

Not enough questions? We learned from speaking with our Customer Experience Team and reading email correspondence with potential customers, that homeowners who fill out our current offer request flow feel Abodewell did not know enough about their property to effectively value their property.

This got us thinking: are we asking the right questions? Are we asking them the right way? Do we allow folks to tell us everything they need to tell us about their house? Are people comfortable sharing information with us?  We know residential real estate transactions are an incredibly personal and emotional process. It seemed people felt they did not have enough opportunity to essentially “sell us” on the home. How can we provide a flow that demonstrates our respect for the homeowner’s relationship to their home and their need to sell it, while effectively gathering the necessary data to produce an accurate and competitive offer on their home.

Time spent: Through analytics and session recordings, we learned that many users fill out the bare minimum and finish the form in under one minute, while others spend up to 1 hour filling it out in incredible detail. This confirmed that we needed to strike the right balance to accommodate folks who want to power through the process quickly but also offer opportunities for people to communicate as much as they want about their home.

PROBLEM – Data Science and Analytics

The more data, the better: This is at least true internally. The more data we capture, the more accurately the Data Science team can develop valuations on homes. Data Science team had certain requests for information, often related to the condition of the home, or any large renovations that would not be public knowledge. On the user end of the flow, we know length and pace of the form is hugely important in completion rate so we worked with the Data Science team to prioritize gathering the highest impact data.

Measuring success: The current structure of the ORF doesn’t allow for detailed analytics tracking. We can’t effectively measure things like drop-off rate as we can only know which page they dropped off on, but not the specific question. This makes it difficult to iterate upon our existing flow.

SOLUTION

Question by question: We decided the new form will be composed of one question per page, allowing us to capture data for individual questions even if customer does not finish the form. This also helps us internally as we can better identify different user types, and funnel them to the appropriate channels. This also allows for easier modification of the form in the future as we can add or subtract questions from the flow without disrupting the rest of a page.

Only ask users the questions we need to ask: We designed the new form to be as dynamic so that each user type gets a tailored experience and only sees questions relevant to their specific home and sale plan. Pre-populate wherever possible and ask users to confirm. Good for when we already have access to data with a relatively high degree of certainty, through publicly available sources like tax data, Zillow, HouseCanary, etc., but need confirmation from the seller.

SOLUTION CONT.

Pace: Incorporate a mix of question types including 1-click single-select, multiple choice, drop-down, to longer free form answers. Allow folks to get through the easy stuff quickly, but provide a space for them to tell us as much as they want about their home where things are less cut-and-dry.

Cosmetics and condition of home: Something that had been missing from our flow was about understanding the condition of a home. We currently only have one question about the overall condition of the home. The condition of a home makes a huge difference for how the Data Science team values a home, what repairs will be needed, and how that effects our holding times. We have thus added more granular questions about the condition of the home where we ask about more expensive features like flooring, walls, bathroom and kitchen.

Question Templates: Improving design → engineering workflow. Worked closely with engineering to define the best workflow for rolling out the new ORF efficiently. We created question templates, and were strict about when new templates could be created in the event of edge cases. This template system helps not only for the initial roll out, but for any additions or modification of questions in the future. Additionally, we worked with engineering to determine how data captured on the front end would be mapped to our internal Admin Dashboard, as this would affect the ORF and of the Admin Dashboard UI.