Shorten the Distance Between Investment, Evidence, and Action
Clicks and conversions may be easy to measure, but they do not always explain whether advertising created additional value or simply captured demand that already existed. According to Michael Shang, SVP of Technologies at StackAdapt, when outcome data enters the active media workflow, planning becomes more iterative.
Topics
Different identifiers, reporting timelines and attribution methods can produce very different views of performance. Clicks and conversions may be easy to measure, but they do not always explain whether advertising created additional value or simply captured demand that already existed.
Media exposure, online activity, transactions and business results often sit across different systems, making it difficult for marketers to connect investment with the outcomes that matter most.
AI is adding another layer to the conversation. These systems can process more signals, identify patterns and support decisions at a speed that would be difficult for a human team to match. But technology does not remove the need for judgment. Marketers still need to define what success means, understand the limitations of the data and challenge systems when the objective or the evidence is wrong.
For Michael Shang, SVP of Technologies at StackAdapt, the next stage of advertising measurement is about moving beyond proxy metrics and channel-level credit towards business outcomes and incrementality.
As AI assistants also change how consumers discover products and make decisions, the way marketers plan, target and measure media will need to evolve alongside those behaviours.
In this interview, Shang discusses the barriers to connecting media investment with business impact, how outcome measurement can influence campaigns in real time, where marketers should trust AI and where they should challenge it, why attribution needs to become a decision system, and why incrementality deserves greater attention as advertising measurement evolves.
Excerpts from the interview;
There is a larger focus on proving ROI. Yet, connecting media investment to actual business impact has remained one of advertising’s biggest challenges. What are the barriers here, and how would you advise marketers to overcome it?
The biggest barrier is fragmentation. Media exposure, online activity, offline transactions, and business results often sit in different systems, with different identifiers and reporting timelines. Marketers end up relying on whichever signals are easiest to access, such as clicks or conversions, even when those signals do not reflect the outcome the business cares about.
Attribution also requires judgment. Transaction data can show that a purchase happened, but marketers still need a credible methodology for determining the role advertising played. That means accounting for factors such as seasonality, purchase cycles, geography, channel overlap, and the difference between correlation and incremental impact.
I would start by defining the business outcome before choosing the media metrics. For a destination marketer, that might be visitor spending, transactions, or origin markets. For a retailer, it could be incremental sales or store visits. Teams should then establish a measurement window, data standards, and attribution methodology that everyone understands.
The final step is making those insights available early enough to act on them. Measurement creates far more value when it informs live budget, audience, geographic, and creative decisions than when it arrives as a retrospective report several weeks after the campaign ends.
Your latest announcement focuses on measuring business outcomes while campaigns are still running. How do you see that changing the way teams plan and optimise in real-time?
It shortens the distance between investment, evidence, and action. Historically, an economic impact study might confirm that a campaign generated value, but only after the budget had already been spent. Teams could apply the findings to the next campaign, but they had lost the opportunity to improve the current one.
When outcome data enters the active media workflow, planning becomes more iterative. Marketers can begin with a clear hypothesis about audiences, markets, channels, and creative, then compare that hypothesis with spending and visitation patterns as the campaign develops. If one origin market is producing greater economic value, or one strategy is generating a stronger return on ad spend, the team can respond while there is still time to shift investment.
That does not mean reacting to every daily movement. Transaction-backed measurement still requires sufficient scale and appropriate attribution windows. Teams need thresholds and guardrails that prevent them from optimising against noise.
The larger change is cultural. Measurement stops being a report owned by an analytics team at the end of a campaign. It becomes part of the operating rhythm shared by media buyers, strategists, analysts, and business leaders.
AI is helping people make decisions faster than ever before. As these systems become more sophisticated, where should teams trust the technology, and where should they still challenge it?
Teams should trust AI with problems where the objective is clear, the feedback loop is measurable, and the system can process more signals than a person reasonably could. Bid optimisation, pacing, forecasting, pattern recognition, and rapid analysis are good examples. AI can continuously evaluate thousands of variables and surface opportunities that a team might otherwise miss.
People still need to challenge the objective, the inputs, and the boundaries. An AI system can optimise very efficiently toward the wrong outcome if the goal has been poorly defined. It can also reproduce gaps or biases in the data it receives. A high confidence score should not be mistaken for a causal explanation.
Human judgment matters most when decisions involve brand safety, privacy, regulation, unusual market conditions, or trade-offs that cannot be represented by a single KPI. Teams also need to ask whether the result makes commercial sense and whether the underlying evidence is strong enough to support the decision.
I see AI as increasing the value of good judgment. It gives marketers more speed and analytical capacity, but people remain accountable for deciding what success means, which constraints matter, and when the system should be questioned.
Do we need to reimagine how attribution data can fuel better decisions on performance and channels? What needs to change in the current reality?
Attribution needs to become a decision system rather than a contest over which channel receives credit. Too much reporting still focuses on assigning a conversion to a touchpoint. That can produce a clean dashboard without answering the more useful question: where should the next dollar go?
The answer will rarely come from one attribution model. Marketers need a layered approach that combines platform reporting with transaction or conversion data, incrementality testing, and broader methods such as marketing mix modelling. These methods answer different questions. Used together, they can reveal both what happened and what was likely caused by the media investment.
We also need greater consistency in how channels are compared. Attribution windows, identity methods, and conversion definitions can vary significantly from one platform to another. A channel should not appear more effective merely because it grades its own performance under more favourable rules.
The current reality is imperfect, so teams should be transparent about assumptions and confidence levels. Better attribution will not eliminate uncertainty. It will make uncertainty visible and give marketers enough credible evidence to make stronger allocation decisions.
You’ve recently been exploring how AI assistants like ChatGPT are changing discovery. As more people rely on AI to discover products and make decisions, how do you see it reshaping media planning, targeting and measurement over the next few years?
AI assistants are becoming task engines. People can describe an objective in natural language, receive a tailored recommendation, compare options, and move much closer to a decision without following the traditional sequence of search results and website visits.
Media planning will need to follow that change in attention. Search will remain important, but marketers will need a broader mix of channels that can create demand before a person asks an AI assistant for an answer. Programmatic, native, connected TV, audio, and digital out-of-home can all help build the familiarity and credibility that influence later discovery.
Targeting will also move beyond keyword capture. First-party data, contextual signals, demonstrated intent, and the environment surrounding a decision will become more important. Brands will need content that AI systems can understand and cite, along with media that reaches people outside the shrinking click path.
Measurement will gain new indicators, including AI referrals, citations, answer visibility, and assisted discovery. Those are useful diagnostic signals, but the business objective remains revenue or sales. Because AI systems are opaque and changing quickly, marketers should use controlled tests, observe the downstream outcomes, and adapt instead of assuming that any emerging visibility metric is automatically valuable.
If you could change one thing about the way advertising success is measured today, what would it be? And what’s the one metric or mindset you believe deserves far more attention over the next few years?
I would stop treating proxy metrics as the final definition of success. Clicks, impressions, completion rates, and engagement can help diagnose a campaign, but they should not become substitutes for the outcome the organisation is funding advertising to achieve.
The most important measure will continue to be a business result: sales, revenue, profit, visitor spending, store visits, or another outcome tied to growth. The exact result depends on the organisation, but it should be defined before the campaign begins and carried through planning, optimisation, and reporting.
The mindset that deserves more attention is incrementality. Marketers should ask how much additional value the investment created and what would have happened without it. That changes the conversation from claiming credit for existing demand to understanding whether advertising produced a meaningful difference.
It also leads to better budget decisions. Average return on ad spend can describe what has already happened, while incremental and marginal return help teams decide where the next dollar is most likely to produce growth.
ALSO READ: Has RCS Solved Mobile Marketing’s Biggest Problem: Trust?
