Brand measurement needs a new source of truth | MarTech

The great, historical battle in the marketing industry is brand versus performance.

Growth-minded companies know long-term revenue requires branding and the development of a trustworthy name. Yet they’re dogged by competition from performance-minded products that seem to spring up overnight with little more than a landing page filled with artificial reviews and stock imagery.

For a hot minute, we were talking about performance branding as if we’d invented New Coke, in a bid to find a one-shot balance between these two forces. Guess how that worked out.

While we all agree that brand and performance are essential for near-term cash and long-term sustenance, what’s up for debate is how we define success on the brand side of this battle. We’ve spent decades relying on surveys and traditional brand metrics to tell us whether our investments are working.

But those measures capture only part of what’s happening between a consumer’s experience with a brand and the decision they ultimately make. Let’s look at what branding outcomes really tell us, how to move beyond information loss to real consumer understanding, and how we’re implementing that in real life.

What are branding outcomes?

First, let’s clear up some terminology. When we measure branding outcomes, we look at awareness, equity, and valuation.

With awareness, we’re looking for familiarity, knowledge, and whether our brand is top of mind when consumers think of a category. For example, if the question is, “When you think of sneaker brands, which ones come to mind?” is our brand, let’s say Nike, included as a response?

When it comes to equity, we need to go a level deeper. What does writing the word “Nike” do to the value of a shoe when it’s written on the side? Valuation takes that equity and puts a total dollar figure on it.

The problem with brand sentiment is human sentiment

For the past several decades, our industry has relied on major measurement firms like Kantar, Ipsos, and NielsenIQ to create these measures of awareness and equity. These giants have millions of data points, sophisticated models, and recently modernized plumbing for today’s complex and nuanced discovery surface.

However, beneath all of that sophistication, the input still boils down to asking a panel of humans what they think about a brand, then using their answers to measure its health.

It’s time to question that as our source of truth.

Why? Because we already know people are unreliable narrators of their behavior. A shopper who says he prefers local businesses has an Amazon order history that says otherwise. Or a traveler will say her top choice for an airline is based on service and comfort while she’s sitting at home, but if she buys a ticket tomorrow, suddenly price becomes the deciding factor.

Or try asking a friend why they prefer Apple, and they’ll probably be able to produce a perfectly rational explanation about usability or privacy, but can they articulate how much of their preference comes from social signaling, prior experience, aesthetics, or exposure to decades of marketing?

Moving beyond information loss into consumer understanding

How do we address this gap when it comes to something as complex as measuring brand equity?

Let’s process map this:

• Marketing happens.
• Consumers experience marketing.
• Perceptions about the brand change (or don’t).
• Researchers ask consumers about those perceptions.
• Consumers do their best to remember and verbalize their answers.
• A model converts answers into brand metrics.

These layers of abstraction between marketing impact and modeling brand metrics are riddled with information loss.

Think about everything that happens between the consumer’s experience with a brand and them checking a box on a survey. It can be a lot.

Before, we had to compress the richness of human behavior and experience because it was impossible to analyze every consumer pattern. But now we can consider other real-world observable behaviors that are happening concurrently, and we can augment our models and create always-on measurements that are validated by real outcomes.

What augmented brand metrics look like in practice

In the last few months, we’ve been building this exact type of augmented model and have started to see meaningful impact on how our clients allocate their budgets and balance acceleration between brand and performance. Here’s what goes into our model:

• Mental availability: What evidence shows that consumers know and consider the brand? Consumer search, traffic, and prompt analysis can indicate if a brand is top of mind or included in the outputs when a customer is in a research or shopping mindset.
• Perception: What do consumers associate with the brand? We use social and ecommerce unstructured interactions that are individually evaluated, contextualized, and streamed for an always-on analysis.
• Commercial power: What commercial premium does the brand command? This is both owned and retailer price elasticity, which answers the question, “How much does your brand have the right to charge while still maintaining critical sales volume?” Even the stock price can be carefully weighed as a vote from investors on its durability and strength.

Let’s also be clear: AI discovery is a new observable surface that we need to fit into our measurement frameworks. Consumer prompts that ask for a product recommendation might come without a brand input from the user. But the AI-generated response will tell us, in a way, what the AI’s mental availability is for a brand.

Agentic brand perception and awareness are critical to shaping what the user considers and, eventually, to driving a direct purchase by the AI.

We use all of these signals indexed against a brand’s competition to ensure we’re looking at the market footprint and not overinflating trending or seasonal factors. If your brand’s perception is moving in a positive direction, we need to ensure it’s specific to you and not a category shift. The signals are then weighted by their impact on the revenue outcome. In other words, the pillars aren’t considered equal. Instead, they are statistically evaluated to determine their ability to predict the business results.

The results we’re seeing from augmented brand metrics

And it’s producing real impact. One of our clients launched their first high-dollar branding campaign against some internal headwinds. The brand lift study we’re running concurrently will only start showing results about a month after launch, but we can monitor, in near real time, how indicator signals like share of search and perception are shifting and how brand resonance is developing.

This means the branding budget can become defensible more quickly, and the team can make investment decisions while there’s still time to affect the outcome. The eventual study results become another validation signal and not the first moment the client learns whether its investment is working.

This is all unlocked because of the broad availability of AI models, which can consume and make sense of huge amounts of heterogeneous, unstructured information at a cost that makes continuous analysis practical.

The future of brand measurement is multisource

I’m not saying that surveys don’t serve a purpose. But they’ve never perfectly represented brand equity. We’ve always known that. They’re impressively good at turning something that was incredibly difficult to observe into something structured. With AI, that’s no longer a constraint.

We’ve spent decades defining brand equity based on the data we could collect, rather than on the data necessary to understand it. It’s time to redefine our source of truth. The future of brand measurement doesn’t need a single source of truth at all.