Why We Transitioned to Hightouch at WeightWatchers

Technology changes quickly, but the underlying use cases and problems that technology solves always stay the same. You can’t fall in love with technology; you have to fall in love with use cases.

At WeightWatchers, our digital advertising efforts are substantial, so even small optimizations to our audiences can have huge implications on our return on ad spend (ROAS) and customer acquisition costs (CAC). To put it into perspective, even a one percent decrease in our cost per acquisition (CPA) translates to huge savings.

For years, we’ve relied on a legacy data onboarding platform to power all of our audiences and data onboarding use cases, but we knew we could get all the benefits using a Composable CDP that ran on top of our existing data infrastructure. In this blog post, we’ll discuss why we transitioned to Hightouch and the changes it has brought to our advertising approach, including better match rates, faster data onboarding, and more accurate audience targeting.

Speed is Everything in Advertising

Founded in 1963, WeightWatchers is one of the world's first and largest global weight management organizations. Our mission is to empower our members to live healthier, longer lives as the trusted authority in evidence-based weight health and community support—for over 60 years strong.

Our business is built around a subscription model, which means the marketing team is focused on acquiring members through various channels at efficient LTV:CAC ratios. But not all things are equal, and this definitely holds true from a member standpoint. Over the years, we’ve learned that net new members are more valuable than existing ones, so a large percentage of our ad spend is focused on acquisition. We've come to realize that two metrics have an outsized impact on our advertising performance: campaign velocity and audience match rate.

Campaign velocity measures the speed at which we can launch and test new campaigns targeting different audiences across our advertising channels. And match rate refers to the number of users that ad platforms can effectively reach or suppress. The higher the match rate you get in an ad platform, the more control you have over who sees (or doesn't see) your ads. Both of these metrics have huge implications on our core metrics like CAC, ROAS, CPA, etc.

The key challenge we needed to solve was giving our marketers the ability to self-serve and build audiences without engineering support–and that was what our data onboarding platform was lacking.

The Problems with Data Onboarding

On paper, data onboarding seems really simple. You upload your audiences to your onboarding solution; the platform then enriches that data and onboards it directly to your ad platforms. Ideally, linking your offline data to your media channels for advertising should be simple. However, the process can be more complex than it initially seems.

Complexities

The primary control you have is over the data you send to the platform. You upload your audience to your data onboarding solution, which then facilitates the matching process between your data and the media platform. However, visibility into this process can be limited, making it challenging to analyze beyond the match rate. Understanding what factors might be affecting your match rate or troubleshooting issues can be difficult without further insights.

Timeliness

To use a data onboarding tool, you first have to get your data into the platform, which means three steps have to take place before you can run any campaigns in your ad platforms–and syncing data to an onboarding solution is usually done via manual uploads or through custom-built data pipelines to automate the process. Our previous process looked like this:

1. Submit a ticket to your IT team to build your audience: In our case it would take upwards of two weeks just for the data team to build those audiences.
2. Upload that audience to our data onboarding tool: Every time we wanted to ingest our data into our previous onboarding tool it had to be manually imported via SFTP, and this process took anywhere from 3-5 days. Note: this also meant that we had to store PII data in the solution as well, which was not ideal from a security standpoint.
3. Onboard your data to your ad platforms: Once you’re ready to send your data to your ad platform, it takes another 3-5 days for it to be processed–and another 1-3 days for the audience to settle in the ad platform.

The challenge we faced was that it took around a month to target a single audience. This can impact effective marketing because relevancy is key. The faster you can build and target audiences, the better feedback loop you can have between your campaigns and their performance. We like to think of this as time-to-learning, or the speed at which you can go from idea, to campaign, to measurement.

Observability

In our experience, our marketing team faced challenges in trusting the data because they couldn’t truly self-serve. Since our marketing team doesn’t read SQL, they were unable to review audience definitions to ensure the data team interpreted their request properly. There was no way for them to understand how often audiences were being updated or which members were in specific audiences. This occasionally led to targeting discrepancies. The lack of a shared vocabulary or user interface between our marketing and data teams made alignment more challenging.

Pricing

When using a data onboarding tool, it’s important to understand the various costs involved. These include platform fees, destination fees, and payment for unique records, which results in a complex pricing structure. In our case, this required a significant investment, both in terms of budget and the allocation of full-time data engineers, to manage the onboarding and enrichment of our data.

Unlocking Speed & Flexibility with Hightouch

We knew there was an opportunity to simplify our architecture and simply leverage all of the work our data team had built in Snowflake to unlock self-serve audiences, greater flexibility, better match rates, and faster data onboarding times—and this realization led us to Hightouch.

Hightouch doesn’t store any data. Instead, it operates on top of our existing data infrastructure. This means:

1. All of our PII data is securely maintained and governed in Snowflake
2. Our marketers can self-serve and access any data point they need at any point to build and target any audience
3. We can control and send over more first-party data to our ad platforms
4. Hightouch enriches all of our data in-flight to the ad platform

Initially, the goal was to simply be at parity with the match rates of our data onboarding platform, but we quickly realized that Hightouch far surpassed everything. We were able to reduce our data onboarding time by two weeks and give our marketing team complete self-serve audience-building capabilities. With Hightouch, we get the same benefits of traditional data onboarding, just with far more speed, better flexibility, and security–and no risk of storing data in another platform.

Here’s a diagram comparing and contrasting the old architecture with the new architecture.

Traditional Data Onboarding vs. Hightouch

Additionally, we’ve observed a significant reduction in vendor and operational costs since adopting Hightouch MatchBooster.

Closing Thoughts

If you’re using a traditional data onboarding platform and you haven’t tested Hightouch Match Booster, it might be worth considering based on your needs. Exploring new options can help ensure you’re not missing out on potential improvements.

If you’re interested in learning more about how we use Hightouch to power our advertising use cases, check out the full case study.