Learning to Lead with Data and Leave Gut Feelings Behind

Data is everywhere, but it’s what you do with the data that counts. 4Sale has always had top-notch data, but the product team used it to reflect on past events rather than predicting future success. We turned things around by adopting the right product analytics platform, which transitioned us from being tech-led to product-led. In the process, we also created a better experience for our users.

4Sale is the largest online marketplace in Kuwait, with more than 6,000 new posts every day. Our mission is to facilitate the best experience between Kuwaiti buyers and sellers in each of our verticals: automotive, property, services, buy & sell, electronics, and jobs. Nothing stays in the system for more than 30 days, and most of the listings are paid, ensuring a consistent stream of fresh, quality content—essential in the listings business. It also means we have a constant supply of fresh, quality data.

Our quest to become more strategic

I’ve been at 4Sale for six years, starting as a mobile developer and becoming a product manager three years ago. For years, we had difficulty answering our most pressing questions, despite all of our data. We used to view everything from a technical perspective. We fixed the bugs that needed fixing, but that wasn’t enough to improve our product. We weren’t focusing enough on the path ahead or taking a data-driven approach to developing new features.

We used different solutions, each one serving a niche purpose. We used one for push notifications and collecting data to target specific customers. We used another app which recorded user sessions we could view at a later time. It helped monitor feature releases and see how to fix bugs, but even with all the appropriate privacy constraints, recording activity in the app still felt creepy.

All the solutions we used shared one thing in common: they all looked exclusively backward instead of forward. The idea was to interpret historical data, and there was no projection about what came next. We wanted to get information about personas and tap into how users interacted with our platform to identify the next steps. That was neither clear nor easy to find using our existing platforms.

Producing a successful product is about more than fixing bugs and examining past behavior. You have to focus on the path ahead.

We were also focused on putting out new releases. We moved rapidly—and it was exciting—but producing a successful product is about more than quick releases. Supporting that pace without a long-term vision isn’t possible, and we couldn’t scale with our existing architecture. To maintain our revenue in the long term, we needed to become a more product-led business, one where product managers could make strategic decisions and develop well-researched product roadmaps supported by behavioral data.

A solution that prioritizes product teams

As we looked at options in the market, we were drawn to Amplitude because it could help us optimize our data and inform us about the future. Other platforms try to be everything to everyone across an organization, but one of the reasons we chose Amplitude over other platforms is because it prioritizes product teams and their needs. As our product team was determining how we measure success for each feature, Amplitude could help us build those product-specific metrics.

It’s great to have strong opinions, but part of becoming product-led is focusing on behavioral data to support those opinions and decisions.

To better harness our data, 4Sale built a data warehouse with the help of the Amplitude Professional Services team, who ensured all the Amplitude data would be in the right place. We have dedicated data scientists with their own reporting systems and mechanisms through Looker, but Amplitude’s team helped us sort out data flows to store everything in one place and give each team what they needed.

Our product team uses Amplitude the most, though other departments use it, too. The marketing team, for example, pushes targeted personalized notifications through a Leanplum integration based on information gleaned from Amplitude’s Behavioral Cohorts. Our C-suite executives look at reports generated through a combination of Amplitude and Looker. We use Amplitude to share the successes of our new features with the rest of the company using quantitative measurements.

How Amplitude helps launch new features and improve existing ones

We look to Amplitude to judge stability, understand how current users behave and interact on the platform, and determine how new features affect the user journey.

We first look at Funnel Analysis charts to see who is dropping off where, or if they’re converting, why. Like any other classifieds marketplace, there are two big funnels we want to optimize: sellers who post listings and buyers who search for items or services. We focus on the number of successful uploads and the number of people who make contact about a particular listing. After determining who drops out of those processes, I’ll use Amplitude’s Journeys feature to dig deep into what these people do next in the application. Do they try to find another way to perform the same action? Another way to post, or another way to search? Journey’s machine learning automatically focuses on the exact moments with the greatest impact.

Of course, our users are not necessarily exclusively buyers or sellers. In fact, most of them are both, and we use cohorts to understand users better. We want to identify how we can convert buyers into sellers, and after they post their first listing, how do we convince them to keep posting? Both of these questions can be answered with Amplitude.

We’ve also leveraged analyses within Amplitude to develop key product changes, one of which is a global search feature we released in Q4 2021. Previously, users had to search for an item within a specific category, instead of across the site. Within Amplitude, we saw that this cumbersome search process led to users dropping off without buying. Our new global search feature allows users to search for a word used anywhere throughout 4Sale. In the first quarter of its release, we saw 40% of users adopt this feature—a huge number considering we have over 1 million users on the app.

We also use Amplitude to enhance our existing functions. If sellers don’t have a good experience with us, they’ll stop posting listings, giving buyers fewer posts to choose from and less reasons to return. Users have to register to pay for their listing, but Amplitude highlighted that potential sellers were dropping off when we prompted them to register. We responded to this drop off by creating a more interactive posting flow with fewer steps. This new flow has resulted in more listings. Without Amplitude, it would’ve been much harder to pinpoint and eliminate that friction point.

Improving the user experience for everyone

Being a market leader is tricky. Your competitors are always looking to take users away, so you have to know your customers better than anyone else. But in a country of 4 million people, we have 1+ million users—making it nearly impossible to generalize about an “average user.” Some users might sell a household item for five Kuwaiti dinars (around 16 USD), while others sell a house for 1 million KD. It’s hard for us to create the same experience for everyone, but Amplitude helps us better understand user personas based on their behavior. That helps us optimize our existing features and successfully build new ones.

It’s hard to create the same experience for all users, but with the right data, you can understand user personas better based on real-time behavior.

As 4Sale continues to scale, our mission remains to facilitate smooth transactions between buyers and sellers. The more we achieve this goal, we might see users spend less time on the app—but only because we’ve made a faster experience. We have learned to leverage the data within Amplitude to take our product to the next level and offer a seamless and efficient user experience.