For the past two decades, we’ve let our success be defined by clicks, impressions, click-through rates, rankings, and everything else that Google would throw our way. Then came AI, and all of a sudden we started paying attention to new metrics like brand citations, prompt coverage, and conversions from AI referrals, among many others that added a whole new measurement layer. And there’s nothing wrong with traditional metrics. In fact, recent analyses of Google’s ranking systems suggest that these signals, along with user interactions, increasingly help search systems infer the value and relevance of content.
The problem is that they’re all measuring the same thing: the outcome of a decision.
They tell us what happened after someone chose a result, but very little about what made that result the right choice in the first place. Or why we are (or aren’t) mentioned in a prompt answer. And in my opinion, this leaves out a massive opportunity hidden in our data.
Because, while explicit choices and behaviors that result from them are great indicators of what works (and, arguably, a predictor of SEO success), everything that happens before a decision is made is equally important: it allows us to identify gaps in our strategy and what we could be doing at scale to influence the likelihood we’re not only visible, but the chosen option among those available.
This is why I’ve started thinking less about traffic and more about what I call “Decision Distance”: the gap between the motivations driving a user’s decision and the messages a brand communicates at each stage of the customer journey.
Why Aren’t Traditional Metrics Enough In The Age Of AI? #
Truth to be told, they were incomplete before, but the introduction of LLMs into search journeys just amplified the blind spot in understanding user behavior beyond what’s immediately available as a result of an action.
Don’t get me wrong, traditional metrics are still great when you’re building out a strategy and raising dev tickets to make sure you’ve got the basics covered, both from a technical and demand coverage perspective; but most of them only capture the outcome of a decision that has already been made, and if we leave it at that, we’re missing out on some of the biggest opportunities to align with our audience, especially now that LLMs are significantly shortening search and decision journeys.
As a matter of fact, the convenience of having an answer synthesized to us before we even visit a website (whether it’s on Google’s own ecosystem with AI overview, or directly on LLMs with conversational queries) has integrated, when not replaced, the systematic research process we used to perform to identify a matching result on search engines. And often, those summarized results cover not only the query, but our specific intent and even temporary states, a kind of personalization that’s hard to let go of. Besides providing convenience, LLMs try to mirror our choices based on what they know about us, our values, our preferences – or those of people like us. For a brand, this makes it harder to rank as an option if we are not perfectly aligned with the search intent, the values, and the emotional states of the person behind the search, since that’s exactly what the LLM is trying to provide.
As AI Overviews and conversational search encourage users to evaluate results before clicking, more of the decision-making process happens outside of our websites, and outside our field of vision.
That’s where traditional SEO metrics begin to show their limits. They tell you what happened, but don’t expand on why, leaving the bigger picture incomplete.
For example, CTRs tell you that people clicked through, but not why a result felt more compelling than another; bounce rate tells us that users left, but not what expectation wasn’t met; the same goes for most of the traditional metrics we’ve been taught to care about. They tend to lag behind, and tell us a story of an action that’s already been taken, rather than what drove a user to that specific outcome. By the time those metrics are measured, the decision (human or agentic) has already been made and cannot be influenced.
So, if decisions are increasingly being shaped before an action is taken, measuring clicks alone tells us less and less about why people chose one result over another. And that’s the gap I believe we should start focusing on.
What You Should Measure Instead: Decision Distance #
Every decision is driven by a combination of functional, emotional, and social drivers. The closer your messaging aligns with those drivers, the more likely someone is to move forward. People rarely convert because of the content itself – they perform an action because what the content conveyed moved them closer to a decision they were already evaluating. So instead of using only clicks, rankings, and traffic as our main key performance indicators, I believe we should also measure the alignment between our messaging and the motivations that actually influence a decision.
I call this “Decision Distance”: the semantic gap between a user’s underlying decision drivers and the messages a brand communicates throughout the customer journey. In other words, Decision Distance estimates how far your messaging and offer are from the motivations that actually move someone closer to an action.
This is the sort of calculation we do every day, almost on autopilot. In our daily lives, we are often making decisions and evaluating the fitness of an option to our needs, goals, and values – and by doing so, we are consistently measuring the gap between what those options offer and what our needs are.
Think about your commute to work, for example. You might not take the fastest route all the time, but choose the option that best fits what matters in that moment, recalibrating and adjusting the journey. Maybe it’s avoiding traffic, stopping by the gym on the way, or getting home as quickly as possible. The same extends to our online behavior: which streaming service to subscribe to, or which software to buy. We constantly evaluate how well an option fits our needs, goals and values before making a choice. And because we are creatures of convenience, our natural tendency is to choose the options that most closely align with our set of implicit requirements, where the distance perceived is the lowest.
Measuring the perceived distance between what your users want and what your messaging provides is the missing piece in understanding why you might be losing people and citations – and, crucially, how to solve it.
How does this look in practice? Imagine someone searching for payroll software. If their biggest concern is trusting a provider with sensitive employee data, but your page spends most of its time talking about features, integrations and dashboards, then your Decision Distance is high, and it’s unlikely to generate a conversion. Your content might have matched the main query, but not the requirements for a decision in your favor.
Decision Distance informs what psychological drivers generated that behavior in the first place, and how well we are catering to them.
That’s why I see it as a complementary measure rather than a replacement for traditional SEO metrics. Rankings, clicks, and impressions tell us whether our content was seen or acted upon. Decision Distance helps explain whether we addressed the motivations that make that action possible in the first place.
How To Measure Decision Distance #
Using sentence embeddings and semantic similarity, you can estimate how closely your offer and messaging are aligned with a library of decision drivers and motivations expressed by your audience. The process I use to find Decision Distance is broken down into four main steps:
1. Identify Decision Drivers From Customer Language
The first step is to identify the motivations behind a user’s search. While queries don’t tell us the full story, they often contain strong signals about the functional, emotional, or social drivers behind a decision, and help us shape an audience’s decision profile.
So define the decision drivers you want to detect and write a short description of each one. These descriptions act as semantic reference points that can be compared against customer language using sentence embeddings, so it’s important you spend some time refining them and reviewing ambiguous classifications. Some of the most common decision drivers are, for example:
Value for money: The desire to maximize perceived value relative to cost.Trust: The need for confidence that the brand will deliver on its promises.Convenience: The desire to minimize effort, complexity and time.Social proof: Confidence derived from the behavior or opinions of others.Quality: The expectation of superior performance, durability or craftsmanship.
And more.
After you’ve defined your drivers, you’ll have to collect your audience dataset.
You can start this exercise with search queries, then expand to social listening, customer interviews, support conversations, or CRM data to build a richer picture of what drives decisions at every stage.
The goal is to map the audience language to the different decision drivers we have selected. I have a Colab script that I’ve been using as an example to talk about this step, and you can experiment with it here, swapping this toy set with your own drivers and audience data.
In my script, I added a short snippet to assign each driver to their own journey stage, so you know exactly where your actions should be (and inform your cross-functional efforts).
2. Measure How Strongly Your Messaging Reflects Those Drivers
Next, analyze your own messaging using the same decision-driver framework. This is where your brand’s drivers profile is estimated, and it’s informed by your own content: product pages, landing pages, blog content, ad copy.
In this step, you map out the brand’s pages, claims, or messaging to the same drivers you have established in Step 1, and your goal is to understand which drivers your brand reinforces, and which it overlooks.
The output can look something like this:
In this example, we see what the top three drivers for each one of these messages are, based on their similarity to the definitions we have set in the previous step.
3. Compare The Profiles To Calculate Decision Distance
This is the real “Decision Distance” calculation step, where you compare the two profiles and isolate the difference between what your content says and what your audience actually needs to hear to move forward. Comparing the two profiles produces an experimental Decision Distance score, alongside the individual driver gaps that explain where the two profiles diverge. Those gaps become opportunities for optimization and prioritization.
Say, for example, that the audience profile reveals quality as a big driver in the awareness stage, but your content leans heavily on convenience instead: the former reflects an unrepresented drive, and an opportunity to better align to what users want to see.
4. Reduce The Gaps Through Messaging And Content Changes
Finally, use those insights to improve alignment. That might mean changing messaging, introducing stronger trust signals, restructuring content, or collaborating with other teams to address needs that aren’t currently being met. This is your roadmap to better alignment with your user.
In this simple example, the audience and brand language show only partial alignment. Audience queries leaned more towards quality and trust, while the brand messaging emphasized social proof, convenience, and safety. Note that these proportions do not represent the percentage of people who care about each driver, but how strongly each driver appeared in the language analyzed.
Traditional search rewarded matching queries. AI increasingly rewards matching the reasons behind them, and understanding those reasons may become one of the biggest competitive advantages in search moving forward.
More Resources:
Rethinking Audience Targeting In A Signal-Loss Era (With The R.E.M. Framework)What AI Overview Click Data Reveals About Consumer Search Behavior: 5 Strategic Insights For CMOsHow Zero-Party & First-Party Data Can Fuel Your Intent-Based SEO Strategy
Featured Image: Roman Samborskyi/Shutterstock