G · DiagnosisExternally proven

Data Monetization: Improve, Wrap, Sell

A model from the MIT Center for Information Systems Research (Barbara Wixom, Jeanne Ross and colleagues): data turns into money along exactly three paths, by improving internal processes, by wrapping the core offering with analytics features, or by selling an information offering of your own. Every path draws on the same five enterprise capabilities but demands them at different levels of maturity. The last step is the hardest: actually collecting the value created and finding it again in the income statement.

When you need this method

You have data, dashboards, and by now an AI budget, yet nobody in the building can say which line of the income statement has moved because of it. Initiatives are justified with potential rather than with realized value, and because the path to money was never fixed, the business unit, data science, and governance are each building something different. What is left is a data estate that carries cost and asserts impact.

Approach

  1. 1For each data initiative, fix which of the three paths you are on: improve internal processes, wrap your own offering with analytics features, or sell an information offering. Name hybrids openly instead of blurring them.
  2. 2Assess the five capabilities for that specific path, namely data asset, data platform, data science, data governance, and customer understanding, and name the weakest one. AI work adds explanation as a sixth, because models are opaque, malleable, probabilistic, and unproven.
  3. 3Give every data asset an owner who carries responsibility across its full lifecycle, the way a product has an owner. Without a named owner, nobody manages quality, access, and reuse.
  4. 4Create access and a shared vocabulary across functional boundaries so that the business unit, data science, legal, and sales are discussing the same thing. Data monetization rarely fails on a single capability.
  5. 5Before you start, fix which financial figure is supposed to move, who measures it, and when it gets reviewed. Skip this step and the value stays an assertion, even when the technology works.

Typical application

A typical B2B SaaS in an industrial setting has collected telemetry from its customers' machines for years, launches an AI initiative on top of it, and twelve months later has a dashboard nobody pays for. The classification is clear: the path is wrapping, not selling. The weakest capability is not the technology but customer understanding, because nobody knows what decision a maintenance planner actually makes in the morning, closely followed by governance, since the legacy contracts do not cover analysis across customers. The team appoints an owner for the telemetry asset, renegotiates permitted use in the contracts, and fixes up front that service call-outs per installation is the target figure and that the service organization measures it. Two quarters later the feature is an argument in renewal negotiations, and the contribution sits in a named place in the income statement rather than in a slide deck.

Limits and counter-indications

The evidence comes overwhelmingly from one research group's own surveys and case studies, is self-reported, and is cross-sectional. The link between strong capabilities and higher returns is therefore well described but not established as causal: successful companies may simply be able to afford strong data capabilities. An independent literature review characterizes the field as nascent, with a thin empirical base especially around pricing information offerings. The model grew up in large corporations with a dedicated data organization; a thirty-person software company cannot staff five capabilities separately, and the team-sport principle presupposes departmental boundaries that do not exist there. On the third path, legal permissibility is not a maturity level but a veto: what contracts or data protection rules do not allow will not become sellable through a better platform. And attributing results to a single income statement line only works cleanly where the finance function can actually produce that line.

How to measure impact

Carry three entries for every data initiative: the chosen path, the named financial figure, and the review date. As a portfolio metric, use the share of initiatives whose value has been demonstrated and located again in the income statement.

Related methods

Sources

  1. 1.Wixom, Someh, Zutavern, Beath: Explanation. A New Enterprise Data Monetization Capability for AI, MIT CISR Working Paper No. 443 (opens in a new tab) · MIT Sloan Center for Information Systems Research / AlixPartners · 2020-07 · academic and scholarly literature · describes the methodCarries the procedure in detail and is freely available in full text: the three paths to money plus the five enterprise capabilities with definitions and building practices, namely data asset, data platform, data science, data governance, and customer understanding. For AI work it introduces explanation as a sixth capability, with four named practices against opacity, bias, probabilistic output, and lack of proven use. Limit: the capability comparison between strong and weak firms rests on self-reported cross-sectional data from 315 companies and shows association, not causation; the explanation section is explicitly labeled a preliminary finding.
  2. 2.Wixom, Ross: How to Monetize Your Data, MIT Sloan Management Review 58(3) (opens in a new tab) · MIT Sloan Management Review · 2017-01-09 · academic and scholarly literature · supports the underlying mechanismThe earliest statement of the three-way split in its present form: improving internal processes and decisions, wrapping information around products and services, and selling information offerings into new and existing markets. It also carries the advice to start with the most promising opportunity rather than pursuing all three at once. Limit: the full text is paid, only the freely accessible opening can be checked, and the capability list is not yet spelled out here.
  3. 3.van der Meulen, Wixom, Beath, MIT CISR Data Research Advisory Board: Mind and Hand. A Decade of Data Monetization Research, Research Briefing XXVI-7 (opens in a new tab) · MIT Sloan Center for Information Systems Research · 2026-07-16 · academic and scholarly literature · evidence of effectivenessCarries the three leadership principles, namely running data assets as products, treating data monetization as a team sport, and realizing the value created with discipline, plus the impact figures: in a global survey of 349 executives, the strongest organizations attributed eleven percent of revenue to data monetization against two percent for the weakest. Limit: these are self-attributed revenue shares from a cross-sectional survey, not audited figures from financial statements, and the direction of the relationship remains open.
  4. 4.Ofulue, Benyoucef: Data monetization. Insights from a technology-enabled literature review and research agenda, Management Review Quarterly (opens in a new tab) · Springer / Management Review Quarterly · 2022-11-28 · academic and scholarly literature · limits the methodAn independent check from outside the MIT research group: this review of 54 peer-reviewed papers from 2013 to 2022 organizes data monetization differently, first by internal versus external and only then into wrapping, bartering, and selling. It rates the field as nascent with a weak empirical foundation, with only a fifth of the papers using case studies, and names pricing for data products and contract design among the open gaps. That tempers any claim that this is a settled management model.

Origin: Wixom / MIT CISR

← Back to the catalog

Work through this method with an AI

ClaudeChatGPTPerplexity