Dec 7, 2023

Brand testing and measurement: How Supermetrics used data and a full-funnel test to develop and launch a new brand identity

15-MINUTE READ | By Edward Ford

Marketing AnalyticsMarketing measurement

[ Updated Jan 24, 2024 ]

How do you measure the impact of brand?

This is a question that has challenged marketers since the dawn of time.

But today, we’ll aim to answer it once and for all.

In this article, I’ll discuss brand testing and measurement and share how we used data to test, develop, and launch a new brand identity at Supermetrics.

Fast forward:

The challenge of brand renewal

For most companies who undergo a brand renewal, it’s a leap of faith.

You develop concepts, pick one—typically after lots of subjective opinions have been heard—and then work tirelessly on renewing all your assets from website and social channels to email and slide templates. When launch day comes, there’s a lot of anxiety, and you end up asking yourself questions such as, will our website traffic drop, are our conversion rates going to fall, and what will happen to our commercial KPIs?

In short, you’re making a bet. And a big one at that.

So, at Supermetrics, we decided to take a different approach to renewing our brand and measuring its impact. We devised a full-funnel brand experiment comprising nine tests across the funnel to let data decide which direction to go.

The evolution of the Supermetrics brand

Supermetrics has grown considerably over the last few years at 40% YoY. As we’ve grown in size, one of the foundational questions we asked ourselves in the marketing team was whether the overall brand reflected where we were as a business—and where we were hoping to go.

We’d had a lot of traction with our approachable yet functional brand offering a simple self-serve purchase flow. But, we’re also an established player in the marketing analytics space, and we saw increased demand from businesses of all sizes, particularly at the larger end of the spectrum with complex data needs and mature analytical capabilities. So, we wondered if the perception we were creating was right for the next chapter of growth.

Remember, a brand isn’t what you say it’s, it’s what they say it is. And we were at a crossroads between staying with our playful and emerging brand or transitioning to a more sophisticated and established position.

The old Supermetrics home page showed a baby-blue hero and a playful 3D figure

These were some strategic objectives we had to consider:

  • We are seen as an established leader in mature markets
  • We saw increased demand from the enterprise segment
  • Trust, reliability, and security are key when working with business-critical data


  • We are seen as a new player in emerging markets
  • Many of our customers are hands-on practitioners where the user is the buyer
  • Offering a seamless self-serve buying motion is an expectation of many customers

There was a strong argument to either stay true to our playful roots or to transition the brand towards something more sophisticated.

But as we say at Supermetrics, “without data, you’re just another person with an opinion,” and it’s fair to say there were lots of opinions.

So, we decided to run an experiment and let the data tell us what the market really thinks. We also wanted to use data the way it should be used, as headlights guiding you forward and not just as a rear-view mirror showing what’s happened in the past.

This is when our full-funnel brand test was born.

5 steps to developing a brand renewal strategy

We’ve developed a brand renewal strategy which includes 5 stages:

  1. Developing two new brand identities
  2. Setting up the full-funnel brand test
  3. Monitoring and analyzing the results
  4. Refining the identity, provisional rollout, and smoke test 
  5. Launching the renewed brand

Developing two new brand identities

So, the test was to discover in which direction we should renew the brand—emerging or established. The first step was to rapidly create two completely new Supermetrics brand identities at either end of the spectrum.

The first identity was a playful and abstract brand with characteristics such as approachable, likeable, curious, helpful, and motivating.

The second identity was more sophisticated and refined with characteristics such as calm, adult, diligent, capable, and trusted.

We split the marketing team into two squads of about three people, with a demand marketer, a performance marketer, and a content marketer in each. Working with our in-house design team, each squad created a new brand identity in about two weeks. For the sake of the experiment, both squads pushed really hard in both directions to see how the market would respond to each identity.

The Playful & Abstract identity

This visual identity embraced the unconventional, using quirky design, abstract shapes and patterns, and playful illustrations.

The Sophisticated & Refined identity

This visual identity maintained a consistent aesthetic across all touchpoints, used one bold color palette, a clean design, and polished, strong language appealing to an enterprise audience. 

Again, all these decisions were made on purpose as part of the experiment to push hard in both directions. Both teams found it difficult, but eventually, we had our two new identities ready for in-market testing.

Setting up the full-funnel brand test

The experiment would consist of nine tests across the funnel to compare which variant, Playful & Abstract or Sophisticated & Refined, would resonate better with our target audience.

The nine tests were split across the funnel and different platforms. We had two top-of-funnel tests, three middle-of-funnel tests, and two bottom-of-funnel tests. In addition, we did a Brand Lift and Ad Recall test on both variants.

  • Top: 6 x 15-second video ad on YouTube (KPI: Views > 50%)
  • Top: 6 x 15-second video ad on Facebook (KPI: Views > 50%)
  • Middle: 6 x 15-second video ad on YouTube (KPI: CTR)
  • Middle: 70 static ads on Facebook (KPI: CTR)
  • Middle: 70 static ads on Google (KPI: CTR)
  • Bottom: 32 static ads on Facebook (KPI: Landing page clicks)
  • Bottom: 32 static ads on Google (KPI: Landing page clicks)
  • Brand Lift test on YouTube 
  • Ad Recall test on Facebook

Our targeting was done on a persona-level within our key geos, and to keep the experiment fair, we created the exact same set of assets that communicated the same message at each funnel stage. The only difference was the visual identity and copy to present these two radically different identities. We built two dedicated landing pages in each brand variant for our bottom-of-funnel tests.

Once we’d built out our landing pages and all the different ads and creatives—which was a whirlwind couple of weeks—we ran the campaigns simultaneously. We invested just over €130k of our annual advertising budget for the experiment since we wanted to ensure we’d reach a statistically significant result for each test in a very short space of time, but you could run a similar test with more or less budget depending on the timeframe, your industry, and overall marketing budget.

To keep track of all the campaigns across the ad platforms and to aggregate the results, we built an automated report in Google Sheets using Supermetrics that kept the test data up to date. This spreadsheet helped monitor the test campaigns and gave us a unified view. Once the test had ended, we used the report to determine the winner and calculate whether the difference in campaign results was statistically significant.

After a few weeks, the results came in, and it was time to reveal the winner of our full-funnel brand test…

The results of the full-funnel brand test

At this point, there was a lot of discussion about which brand would come out on top. Everyone had their own guess, but the results were seriously conclusive.

In short, it was a landslide. One variant clearly came out on top and showed it resonated strongly with our target audience, and it was the Sophisticated & Refined variant.

The Sophisticated & Refined identity won seven out of nine tests (all statistically significant), one was in favor of Playful & Abstract (not statistically significant), and one was a no result.

Results: Part 1

Results: Part 2

Now, while this wasn’t a fully conclusive test at this point, we had a clear signal from the market to move to a more sophisticated brand identity. We also compared the results of this experiment to our existing baseline, which was also a positive uplift, and so we agreed to keep going forward in this direction and move to the next phase of our experiment, which was our smoke test.

Refining the identity, provisional rollout, and smoke test

So we had a clear signal from the market, and at this point, we went gung-ho to develop the new identity. However, since we pushed so hard in the sophisticated direction (remember the enterprise vocab), we pulled the brand back to something a little more authentic and Supermetrics.

Secondly, since this brand renewal meant a website rebuild, it would impact our entire top-of-commercial funnel. As a result, we wanted to be very diligent and structured in how we launched the new identity and site, which is why we decided to run a smoke test.

For the smoke test, we built a microsite in our new identity consisting of 5-10 of our most popular pages and decided to run it in Canada. Here, we simply ran 50% of paid traffic to the new microsite and 50% to our actual site. There were eight major KPIs we wanted to look at, covering acquisition, engagement, and conversion.

Once again, the numbers were conclusive across the board. We saw our test microsite with the more sophisticated and refined identity, which we called our MVP, outperform our existing site. We also received positive qualitative feedback from inside and outside the organization, with one senior product manager saying, “It’s like we’ve grown up overnight.”

Another positive sign we found is that we saw a significant uplift in trials per user on the microsite. There is a strong correlation between trials and revenue, so we were very confident this transition would inflate our entire commercial funnel.

A/B test for advertising traffic in Canada

And then as a precaution, we decided to extend our smoke test to a bigger market, and we ran it in the UK. Again, we saw similar performance as in Canada, so at this point, we were very confident in the new brand identity we were moving towards.

Launching the renewed brand

At the start of the article, I wrote about the anxieties marketers face when rolling out a new brand and site—I know, I’ve been part of several—but after our full-funnel brand test, we accomplished what we set out to do. We were able to choose the right brand identity, measure the impact of the new brand, and forecast the impact it should have on our commercial funnel and business performance. What we were looking for with this test was a direction to go, and that’s what we got.

Being a marketer is terrifying at times, especially if you’re a CMO, which is why you need to lean on data as your trusted advisor. This experiment gave us all the confidence, knowing we listened to the market and based our decisions on what the data told us. It also eliminated the discussions on what color our CTAs should be or which font we should use since the data gave a clear answer.

After several rounds of iteration, we finally landed on the current brand identity, as you can see on our website, which extends right through our product portfolio to deliver a seamless and consistent brand experience across all touchpoints.

We’ve even launched our brand center, where you can learn more about our story and brand. If you’re a partner looking to create content about Supermetrics, you’ll find relevant information about our logos, typography, and visual elements there.

Final word: Yes, you can measure the impact of brand

At Supermetrics, we talk about turning your marketing data into opportunity and using data as headlights to make better decisions. This is one example of how being data and insight-driven as a marketing team can help you uncover growth opportunities you might have missed. Without data, you’re leaving revenue to chance and quite possibly leaving money on the table.

So consider this next time you plan on renewing your brand or undertaking any bigger marketing initiatives. You can rapidly test things and use the data to make better decisions.

And hey, one other thing.

I guess we marketers can finally say, “Yes, you can measure the impact of brand.”

Check out some more useful resources on brand measurement:

Yes, you can measure brand!
Want to learn more about this brand experiment? Listen to this episode of the Marketing Intelligence Show, where we cover the behind the scenes of Supermetrics' brand renewal.
Have a listen

About the author

author profile image

Edward Ford

Edward Ford is Head of Global Demand Gen at Supermetrics and was one of the people leading this brand test. He has been working in marketing and with marketing data for 15 years, loves running experiments with our Growth team, and ensures our marketing is driving pipeline and revenue for Supermetrics.

Turn your marketing data into opportunity

We streamline your marketing data so you can focus on the insights.

Book Demo