Re-inventing Yelp Home Services

How I helped transform a home services marketplace that enables consumers to confidently hire great pros.

In the last few years of the 2010's, an unintentional transformation was taking place at Yelp. We (probably like you) had always throught of Yelp as “the place you go to restaurant reviews”, but our revenue was no longer driven by restaurants - the majority came from home services (plumbers, mechanics, cleaners, etc).

The bad news? Despite it being the largest revenue source, our product was terrible at helping consumers hire great local professionals. It was still built overwhelmingly focusing on restaurants and their customers - and the user journey for someone looking for pizza is very different for someone looking to remodel their kitchen. This was reflected in our numbers - most sessions ended with a funnel drop-off or incomplete hire.

The good news? This was Yelp's no. 1 strategic focus, and it felt like there was a ton of opportunity.

In 2018 I stepped up to lead the newly formed Home Services design team of eight designers. This is the story of how we developed this new vertical within an existing product.

At a glance

Head of Product Design, Services Marketplace, reporting to VP of Design
Responsible for both sides of the marketplace (professionals and consumers) within verticals such as contractors, mechanics, plumbers, etc
Owned the entire services flow - search, discovery, project management, messaging, etc
8 direct reports, two teams with a Lead. One dedicated dotted-line UX researcher.

Just some of the work that came out of the Services Design team during my time as Design Director

Building trust with quick wins

From the outset, it was clear this was going to be a big challenge. As is often the case, my newly inherited design team wanted to make some major changes. They could tell that this hacky restaurant flow was not what our home services customers needed, and they wanted to start over. I sort of agreed with them.

It was tempting to run a big rethink and redesign process, but, Home Services was driving most of Yelp’s revenue - from our leadership teams perspective, they just wanted it fixed, and now.

So, I had a tough conversation with the team where I convinced them that in order for us to “do this properly” we needed to get some wins on the board first. It wasn't a popular choice, but they gave me a shot. Our research team was fully booked, so, with only light guidance we hacked together some simple discovery activities

User interviews: we manually reached out to some pros and consumers and had a chat about their experiences
Usability testing: we got those same people to go through our product and watched them go through our flows
Competitor review: we broke down several of our “best in class” competitors to understand how they were approaching the space.
Messaging analysis: in addition to the funnel analysis, we pulled the logs of all the messages pros had trying to book a job. We went through hundreds of messages and classified them into buckets based on the problems they were having.

This research allowed us to quickly identify a large number of quick wins and small features that we could quickly get into production. At first, we focussed on small UX fixes to help people get through our project submission funnel, increasing conversion. We built on these successes to experiment with and launch larger improvements to the overall experience.

By working our way up from quick wins to larger experience updates we had successfully increased our project submissions by multiple % points and drove significant revenue increases for our bottom line - all the while improving the experience.

And perhaps even more importantly, we were building our reputation as partners that really understood the business. This was going to be critical in what was to come…

Price Guides were a feature that came out of our research - people wanted to know how much a job would cost before reaching out to pros. Our Price Guide landing pages helped them with this information and led to significant SEO-based growth wins.

When the “quick wins” run out

We had a problem - after about a year of these improvements, our wins were slowing down. The quick wins were getting less quick and less winn...ier. We were seeing the limits of relying on our “hacky research” and anecdotal experience:

We didn’t know what pain points existed throughout the entire user lifecycle. Unlike a restaurant booking you don’t dream, plan and hire a contractor for a remodel in one session.
We didn’t know enough about our user's specific use cases. Organising an oil change is different from a weekly cleaner and different from a kitchen remodel).
We didn't understand the business owner experience enough. Many of us had had home services work done, but none of us actually ran a home services business.

At this point, the design team had built credibility with the rest of the business which made it easier to get funding for two larger research projects. In partnership with our research team, we launched Yelp’s first-ever diary study and a series of use case-specific deep dives.

What the research uncovered

Despite all our work, there were still critical problems with our product. Just as my design team had suspected at the start, many of these issues were deeply embedded in our business model which tried to balance advertiser needs with consumer needs and unfortunately, often didn't get that balance quite right:

“I don’t get any jobs or “real” hires from Yelp - these people feel fake!”

— Typical business owner sentiment

“These are a bunch of irrelevant businesses! And they never reply!”

— Typical consumer sentiment

The entire organisation agreed this was a problem, and yet no-one seemed to know what to do about it. Product stakeholders were stuck in the day-to-day growth and no-one agreed on the path forward. We were stuck, but I had an idea: Design Thinking was about to move to the board room.

Aligning leadership towards a common vision

All this time, work was continuing to happen, features were launching, but the impact was slowing down by the week. So it is testament to my team and my cross functional partners that when I wanted to get them together for two days for a workshop to align on our future direction, there was a strong appetite.

We had done all our homework before the workshop - we took all the research, the competitor analysis, the internal feedback, metrics breakdown and usage analysis and more, and gathered for two days.

It wouldn't be a design case study without post-its and sketches. I planned and facilitated the workshop broadly following the Google Sprint process. I even won our senior executives over to design workshopping - it was very fun to see seasoned engineering and business leaders posting stickies and sketching.

The output was an aligned vision of where we wanted the product to be, with a mission statement and supporting documents. The design team then went and mocked this up to create our “north star”

The workshop was incredibly well received but it wasn’t a silver bullet. Whilst it created some inspirational assets and helped us all move in the same direction it also raised a lot of questions - people would look at our north star and ask “Is that even realistic? How long will that take? What will it break?”.

Snapshots from our “design north star” presentation that explored where we wanted to be in the next few years. Features included a rethought landing and search experience and a way to quickly talk to a human to get help.

Design leading a change

Based on the research and visioning work we were able to combine the best ideas into a testable prototype. The core of the idea was a separate sub-brand where we would only have our best businesses that would pay to take part. While this was not dissimilar to some of our competitors, the idea was very controversial for Yelp which was built on the idea of a directory of every business.

We looked at many competitors to study best-in-class UX and visual design for our new sub-brand - some of our crazier early ideas explored the ideas of diverging from the master brand colours - for a while everyone was talking about “those purple mocks”.

Validating our intuition

Facing skepticism, we went out to do some user testing to validate both consumers and pros sentiments around the idea.It was perhaps the most nervous I’ve ever been waiting for research results…and it was a huge sigh of relief when we discovered - users were excited by the new product and didn't have any of the concerns we were worried about

This still feels like Yelp, but it's clearly different to the restaurants' stuff… I like that it's only top-rated businesses…and don't mind that they are paying to take part.

Kathryn, 32, Boston

They didn’t care that the “directory” format was gone and they loved the new UX, but most importantly, the idea of a smaller set of businesses that pay was not an issue for them - as long as they were good businesses. This was incredibly validating, and thanks to the thoroughness of the research team, convinced the entire organisation that there was something in this idea worth testing.

The research gave us the ammunition we needed to convince the CPO and CEO to fund a new cross-functional team to go build it out - we had about 25 people working on this, and it included an engineering team, several PMs, biz ops and data science. It was easy to create the team, as we’d create a lot of “buzz” about the project, so we had lots of volunteers.

The Outcome

After the Yelp Pros team got spun up I assigned a Design Lead and stepped back from the day-to-day. The team launched several experiments to validate our hypotheses and confirm the modelling done by the Biz Ops team.

While the experiments were positive, at some point experiments can only tell you so much and ultimately leadership decided not to take the “big bet” and go all in on Yelp Pros.

However, the project did create lasting change. One of the key features of Yelp Pros did end up launching - it’s called “Yelp Guaranteed” and it’s a way for users to feel safer using Yelp for a large service purchase.

Many of the ideas in Yelp Pros were shipped back into the core product in the form of larger UX overhauls that would not have happened without this initiative, which combined with the continued work by the rest of the Services team contributed to the Services experience being totally overhauled by the time I moved off the team in 2022:

One of our largest projects was our Project Hub. Previously the process of messaging and hiring pros was jerry-rigged on the back of our messaging feature. We built an entirely new project management dashboard for pros and consumers to coordinate the their job.

We reworked the services search experience to capture intent at the start of the flow. This lead to better consumer - pro matching and more hires.

Leaving a review is totally different for a services business than a restaurant - we adapted our reviews flow which led to a significant increase in review conversion.

Perhaps more importantly to me, the real value of the initiative was that it gave the entire organisation the inspiration that we can take “bigger swings” and get excited about an area that was starting to slow down. It brought together product management, engineering, design, data science and biz ops to work together on a larger goal that was initiated by design - we didn’t just have the seat at the table, we had the head chair.

What I learnt

The easiest way to drive a team in a singular direction is to use our super-powers as designers - our empathy and ability to talk and understand our users, our tools for collaboration like workshops, and our ability to visualise a world different to how it is today and to tell stories and get people excited about a future that doesn’t exist yet - but could.

But that said - it's not enough just to collaborate, extract and distil others' ideas. That’s still super important - but you also need to take it a step further and form your own point of view on how your product, team and company need to change. It’s not enough to know the problems and have a flashy vision and sit back and let others do the work - you need to be able to say “Let’s go in X direction - and here's how we might do that”.

Empathy, collaboration, storytelling and having a point of view are all things I now teach and expect from my design leaders. By the time I left Yelp, I had distilled it into the design culture of my teams and I am proud of the continued impact those team members are having on Yelp's business, its product, and its users.