What marketing measurement needs to become in the AI era

Here is where we think the category is going, and how we are building for it.

Lifesight is included in The Marketing Measurement And Optimization Platforms Landscape, Q3 2026, by Forrester. It is a good moment to talk about where this category is headed, because in our view it is changing faster than at any point since we started the company.

The old job was explaining the past

For most of the last decade, measurement was a retrospective discipline. A team would refresh a model once a quarter, build a deck, and present a set of numbers describing what had already happened. By the time anyone acted on the findings, the market had usually moved.

That model had a deeper flaw too. Most of the numbers marketers relied on day to day came from the ad platforms themselves, and every platform is built to claim credit. Add up what each channel says it drove and the total routinely lands at two or three times actual sales. Worse, the inflation is uneven. A channel reporting the strongest return can easily be the least efficient once you strip out the conversions that would have happened anyway. Teams following those numbers in good faith end up moving budget in exactly the wrong direction.

The result is a pattern we see constantly. A meaningful share of spend misallocated, acquisition costs climbing, growth stalling, and a dashboard that looks healthy the entire time.

The new job is running the business

Three things have changed at once.

Budgets now move weekly. Planning cycles that used to run quarterly have compressed, and a marketing leader who waits for the next model refresh is making decisions on stale information.

The channels have changed. Connected TV, retail media and offline now carry serious budget, and none of them produce the clicks that traditional attribution depends on. A measurement approach built around clicks simply cannot see them.

And AI has collapsed the time between a question and an answer. When a team can ask what is driving a shift in performance and get a clear response in minutes, the quarterly deck stops being useful.

What marketers need now is measurement that runs continuously, tells them what their spend actually caused, and helps them act while the answer is still true. In our opinion that is a different product from what this category was built to deliver, and it is what we have spent the last several years building.

What we are building

One causal system. Incrementality testing, marketing mix modeling and calibrated attribution each answer a different question, and each has blind spots on its own. Inside Lifesight they work as a single system. Experiments establish what is genuinely causal. Marketing mix models carry that truth into strategic allocation across every channel, including the ones with no clicks. Attribution inherits the causal read so that daily campaign decisions rest on the same foundation as the annual plan. Marketers get one answer instead of three competing ones and a meeting to reconcile them.

Agentic measurement. The bottleneck in measurement was never the math. It was the distance between an insight and an action. A model would surface that connected TV was lifting branded search, and six weeks later someone would finally move the budget. Our agentic workflows close that gap. They watch continuously for the shifts that matter, explain what is driving them, and turn the finding into a recommendation a team can act on the same day. Every recommendation traces back to causal logic the marketer can see and question.

Open methodology. As more marketing decisions get automated, the models underneath them have to be open to scrutiny. Our research is reviewed by an independent Scientific Advisory Council, whose inaugural member is Dr. Ron Berman of The Wharton School, and which sits outside our product organization. We are committed to publishing the methodology behind our models so that anyone can examine how our numbers are produced.

Skills for the people using it. Better tools only matter if teams know how to operate them. The Foundations, the first level of our Unified Marketing Measurement Academy, teaches causal and agentic measurement from first principles. It is free and self paced, and it is designed for marketers without a data science background.

Why this matters now

Marketing leaders are under more pressure than ever to show what their spend produces, and the tools they have relied on were not built for the questions they are being asked. The shift underway is from measurement as a report to measurement as an operating system for marketing decisions.

That shift raises the bar in two directions at once. The systems have to be faster, because decisions move faster. And they have to be more transparent, because more of those decisions are being made with the help of automation. Speed without a causal foundation just produces confident mistakes more quickly.

We believe the platforms that matter over the next few years will be the ones that get both right. That is what we are building at Lifesight.

Read the report. Available to Forrester subscribers or for purchase.