Explanation over reporting — why analytics is due for a rewrite

For fifteen years the analytics industry optimized for measurement. The next decade belongs to systems that explain. Here is the evidence, and what Veinera is building in response.

Analytics has become extraordinarily good at measurement. It is still remarkably bad at explanation.

Every organization you might care about has dashboards. They have warehouses. They have BI teams. They have more observability into what is happening in their business than any generation of operators before them. And yet, when outcomes miss, the most common artifact produced by that entire stack is a thread that ends in "we're still investigating."

This is not a tooling complaint. It is a category problem — and the numbers make it hard to ignore.

The measurement decade, by the numbers

The scale of investment in marketing analytics over the last fifteen years is genuinely remarkable.

In 2011, the global marketing technology landscape counted roughly 150 products. By 2024, that number had reached 14,106 — an increase of over 9,300% in thirteen years, growing at a compound annual rate of 41.8%, according to Chief Martec's annual landscape analysis. Global martech spending reached $148 billion in 2024 and is projected to surpass $215 billion annually by 2027, according to Forrester's Global Martech Software Forecast.

That is an industry that has compounded relentlessly for over a decade on a single organizing premise: that giving organizations better visibility into performance will produce better decisions.

The results are more complicated than that premise suggests.

What all that investment actually produced

McKinsey's research on martech maturity, published in 2025, offers a frank assessment of where that investment has landed. Their finding: martech's promise remains largely unfulfilled. Most B2C organizations they surveyed self-identified as either "developing" or "operational" in maturity — meaning basic siloed execution, manual workflows, or limited automation. When researchers conducted in-depth interviews with the same respondents, actual maturity ratings dropped further still.

The gap between what the tools can do and what organizations actually get from them is structural, not incidental. Forrester's own survey found that 47% of marketers want to reduce the number of solutions they already use — not add more. Gartner's data shows that marketers, on average, use just 42% of their martech stack's capabilities.

Visibility improved. Understanding did not keep pace.

Accenture's research puts a sharper number on the outcome side: only 32% of business executives say they can create measurable value from data, and just 27% report that their analytics projects produce actionable insights. Only 6% of companies have reached what researchers describe as a mature, insights-driven culture.

The tools multiplied. The explanations did not.

Why measurement alone hits a ceiling

There is a structural reason that more measurement does not automatically produce better decisions. Measurement is designed to surface outputs. Explanation requires understanding the dynamics that produced those outputs — which is a different kind of work.

When a campaign underperforms, the measurement layer tells you the result. It does not tell you whether the cause was a content signal that resonated briefly and then decayed, a distribution pattern that reached the wrong segments at the right moment, a conversion step where friction compounded below the funnel chart's line of sight, or an execution gap in timing or coordination that no dashboard category was ever built to capture.

These are not the same problem. And optimizing against the wrong diagnosis does not fix the underlying issue — it usually compounds it.

The market data reflects this tension directly. While descriptive analytics (reporting what happened) still holds the largest market share among analytics types, prescriptive analytics (recommending what to do next, based on causal understanding) is projected to grow at a 32.72% CAGR through 2031, the fastest of any analytics category, according to Mordor Intelligence's 2025 market analysis. The direction of travel is clear: organizations are increasingly seeking systems that move from observation to direction.

AtScale's 2025 Semantic Layer Summit, which brought together data leaders from major enterprises, surfaced the same theme from practitioners: in the age of AI, explanations matter. If a system recommends an action, decision-makers need to understand how it arrived there — not as a regulatory requirement, but as a prerequisite for trust and therefore adoption.

What explanation actually requires

Explanation is not a dashboard with more filters. It is not a chatbot placed over a data warehouse. It is a structural reinterpretation of what performance data is for.

An explanation-first system is built around three commitments that are meaningfully different from how measurement systems are designed:

1. Treat performance as a system, not a collection of metrics. Content, distribution, conversion, and execution interact continuously. A content signal that works well in one distribution context may perform poorly in another. A conversion rate that looks healthy may be masking a distribution problem upstream. Any layer examined in isolation produces a partial picture — and partial pictures produce confident misdirection.

2. Surface drivers, not outputs. The questions that matter for decision-making are not how much or how fast. They are why the outcome emerged, and what changes if we act differently. A system designed to surface drivers starts from the business question and builds interpretation from there, rather than starting from measurement and working backward.

3. Translate interpretation into direction. An explanation that does not change what the team does next has produced no value. This is the most frequently missed step. Research from MIT Sloan Management Review, based on a survey of 2,719 managers across organizations worldwide, found that the number one barrier to creating value from analytics was not data management or modeling complexity — it was translating analytical results into business decisions. The explanation must resolve into a direction, not a summary.

Why now

Several conditions have converged to make this the right moment to build for explanation rather than measurement.

First, the data foundation is now largely in place. Organizations that have spent a decade building warehouses, pipelines, and integration layers have created the raw material that an explanation layer requires. The constraint is no longer access to data — it is structured reasoning across it.

Second, AI infrastructure has matured enough to support causal analysis at the speed that business decisions require. Methods like geographic difference-in-differences, Bayesian structural time series, and behavioral attribution — which previously required specialist teams and weeks of analysis — can now be embedded into operational systems that surface results in minutes rather than months.

Third, the market is signaling diminishing returns from measurement investment. Martech's share of marketing budgets fell to its lowest level in a decade in 2024, according to Gartner's CMO Spend Survey, even as absolute martech spending continued to grow. The implication: organizations are not abandoning analytics — they are becoming more selective about what they expect it to do.

What Veinera is building

Veinera is being built as an explanation layer, starting with campaign performance.

Campaign performance is where the gap between measurement and explanation is most expensive: decisions are made at speed, against outcomes — offline sell-through, retailer demand, real revenue — that the standard measurement stack cannot observe. The signals across content, distribution, conversion, and execution are rich. The methods for examining those signals causally, rather than correlationally, exist. What has been missing is a system designed to apply them at the pace and scale that operating teams require.

The approach is not to replace reporting. Reporting has a job, and it does that job reasonably well. The approach is to build the layer that picks up where reporting stops: structured interpretation of why the numbers moved, what is driving the outcome beneath the surface, and where to act next.

That layer, built correctly on campaign performance, is the early foundation for something broader: behavioral intelligence as an organizational capability — applied not just to campaigns, but to distribution decisions, demand planning, product development, and the broader commercial environment where consumer behavior shapes outcomes.

This is a longer bet than most software companies take. It is also, we think, the bet worth taking.

---

Sources and references

• Chief Martec. 2024 Marketing Technology Landscape Supergraphic. chiefmartec.com, May 2024.
• Forrester Research. Global Martech Software Forecast, 2023 to 2027. January 2024.
• McKinsey & Company. Rewiring Martech: From Cost Center to Growth Engine. Growth, Marketing & Sales Practice, 2025. Based on survey of 200+ Fortune 500 executives.
• Gartner. 2024 CMO Spend Survey. Survey of 395 CMOs and marketing leaders, North America and Northern and Western Europe.
• Gartner. Statistic on martech stack utilization (42%). Via Edge Linking, 2024.
• Accenture. Research on data value creation and insights-driven culture. Via SR Analytics, 2025.
• Mordor Intelligence. Data Analytics Market Size, Share Analysis and Forecast Report to 2031. 2025.
• AtScale. The Future of Business Intelligence: Trends to Watch in 2025 and Beyond. Including findings from 2025 Semantic Layer Summit.
• MIT Sloan Management Review / SAS Institute. Minding the Analytics Gap. Survey of 2,719 managers across organizations worldwide.