The Measurement of Marketing and Sales, the Interpretation Layer, and Why Knowing Whether Growth Work Has Become the Threshold for Acting on Artificial Intelligence

A research brief prepared for institute fellows and MBA faculty

Over50Pros · Strategic Operations Advisory

Working paper · June 2026 · Fourth in a series — companion to “Who Owns Growth,” “Where Growth Lives,” and “What the Firm Remembers.”

Framing note. This is a research brief, not a sales document. It is the fourth in a series examining the constructs underneath a single operational-maturity instrument, the OQ Score (Operating Quotient), produced within the Business Operating Intelligence (BOI) System. The first brief examined who owns growth (the Team dimension); the second, where the acquisition-and-retention method lives (the Process dimension); the third, where the customer record itself lives (the Tools dimension). This brief examines the fourth and loop-closing construct — whether a business can tell if its growth engine is working at all, and what it does with the answer. This is the Insight dimension of the Grow Work pillar. The author has a direct interest in the instrument under discussion; see §8. The instrument’s constructs are evaluated against independent and primary sources, and every claim that rests on vendor- or consultancy-reported rather than peer-reviewed evidence is flagged as such.

1. The question this brief answers

The OQ Score asks the respondent to select the behaviorally anchored state that best describes their business today:

If asked “Is our marketing and sales working?” we can answer with…A feeling, based on whether the phone rang this month.Top-line revenue numbers only — we can’t separate what worked from what didn’t.A few tracked metrics (leads, calls, proposals) reviewed inconsistently.A defined short list of metrics reviewed on a fixed cadence.A defined short list of metrics the team can interpret and act on without the owner.

Read literally, the item measures one thing: the evidentiary basis on which a firm answers a single question about its own growth engine. Read carefully, it measures four things the management literature has studied separately for half a century. This brief un-compresses them, asks what the evidence says about each, and then asks the question this institute exists to answer: does the arrival of cheap, abundant artificial intelligence make a firm’s own measurement capability more important to growth, or less? The short answer, defended below, is more — and, as in the prior three briefs, the mechanism is structural rather than sentimental.

2. Why “is it working?” is genuinely hard to measure

Three measurement problems recur, and naming them disciplines every claim that follows.

• Self-report bias. The same respondent reports both the state of measurement and, implicitly, their own competence at running the function. A felt sense that “marketing is doing fine” registers as knowledge when it is closer to a mood. This is the well-documented problem of common-method bias, and it inflates self-rated maturity in predictable directions.
• Ordinality. The five anchors are ordered categories, not equal-interval measures. The distance from “a feeling” to “revenue only” is not the same kind of distance as from “fixed cadence” to “the team acts without the owner.” Averaging such levels into decimals is useful but implies more precision than the scale strictly possesses.
• The construct is partly interpretive. Unlike “do you have a CRM,” whether a firm can tell if marketing is working depends on an act of judgment the data does not perform on its own. As we will see, this is exactly the property that makes the top of the ladder hard to fake — and hard for an automated system to occupy.

3. The five anchors are four constructs

The ascending ladder compresses four distinct constructs. Un-compressed, each maps to a separate and well-developed research literature.

Transition	Construct being measured	Research lineage
L1 → L2	Evidentiary basis: intuition vs. evidence of any kind	Availability heuristic [6]; evidence-based management [7]
L2 → L3	Decomposition: an aggregate outcome vs. attributable signals	Lagging-indicator critique; the balanced scorecard [8]
L3 → L4	Cadence: sporadic looking vs. a disciplined measurement system	Management control systems [9]; measurement ability [1]
L4 → L5	Distributed actionability: reviewed vs. interpreted and acted upon, without the owner	Sensemaking [10]; data-driven decision-making [3][4]; the closed loop

4. The evidence, construct by construct

4.1  From feeling to evidence (Level 1 → 2)

The Level-1 firm answers with “a feeling, based on whether the phone rang.” This is not the absence of a method; it is a specific and predictably biased one. The availability heuristic [6] holds that people judge the frequency or importance of an event by how easily instances come to mind. A phone that rang this week is vivid and recent; a quiet channel that historically produces the firm’s best-margin work is neither. The owner’s felt sense of “how marketing is doing” is therefore systematically weighted toward the salient and the recent, not the consequential. Evidence-based management [7] names the organizational version of this failure: decisions made on what leaders believe to be true rather than on what the evidence shows with the predictable result that confident intuition and actual performance diverge most exactly where the leader feels most certain. The move from Level 1 to Level 2 is the move from intuition to evidence of any kind. It is the smallest step on the ladder and, for a firm running on the owner’s instincts, often the hardest, because it requires the owner to treat their own felt sense as a hypothesis rather than a reading.

4.2  From aggregate to attributable (Level 2 → 3)

The Level-2 firm has evidence — but only top-line revenue, “and we can’t separate what worked from what didn’t.” This is the attribution problem, and it is the most common place for otherwise data-literate firms to stall. Revenue is a lagging, aggregate outcome: it arrives after the fact and places every cause into a single number. The lament long attributed to the merchant John Wanamaker, that half his advertising was wasted, and he did not know which half, is the Level-2 condition exactly. Kaplan and Norton’s case for the balanced scorecard [8] rests on this insight: financial measures alone are lagging indicators that report the score after the game is over, and a firm steering by them is steering by its own wake. The move to Level 3 is the move from one aggregate outcome to a few decomposed signals, such as leads, calls, and proposals that sit upstream of revenue and can, in principle, be traced to specific activity. The firm gains the ability to ask not merely whether revenue moved, but which input moved it.

4.3  From sporadic to cadenced (Level 3 → 4)

The Level-3 firm tracks “a few metrics, reviewed inconsistently.” The metrics exist; the discipline does not. This is the distinction the management-control literature draws between possessing a measure and operating a measurement system. Simons’ framework of diagnostic and interactive control [9] treats a metric as a control instrument only when it is reviewed on a regular basis against an expectation; a number looked at when someone happens to remember is not a control but a curiosity. The marketing literature makes the same point at the level of the firm: it is the presence of a marketing performance measurement ability, a defined, regularly used system, not the mere existence of marketing data, that O’Sullivan and Abela [1] associate with higher firm performance and, notably, with higher chief-executive satisfaction with the marketing function. The move to Level 4, “a defined short list reviewed on a fixed cadence”, is the installation of rhythm. It is what converts scattered numbers into a system.

4.4  The hinge: a defined short list, and the difference between measuring and measuring well

Two words in the Level-4 anchor carry more weight than they appear to: defined and short. The performance-measurement literature’s most uncomfortable finding is that cadence applied to the wrong metrics is motion, not progress. Ittner and Larcker’s field study [5] of more than sixty firms and nearly three hundred executives found that most companies measuring nonfinancial performance never establish, and rarely test, whether the things they measure actually drive the outcomes they care about: only about a quarter built and verified a causal model linking their measures to results, and more than seventy percent relied on metrics lacking demonstrated validity. The firms that did the harder work, confirming that a measure connects to an outcome, earned materially higher returns on assets and equity than those that did not [5]. The same study supplies the cautionary image for metric proliferation: an executive dashboard carrying nearly three hundred measures, which is not insight but its opposite — the noise in which insight drowns. This is why the anchor specifies a short list. The discipline being measured at Level 4 is not “track more”; it is the rarer discipline of choosing the few metrics that are both valid and decision-relevant, and refusing the rest.

4.5  From measured to acted upon, without the owner (Level 4 → 5)

The top of the ladder adds a single decisive clause: the short list is one “the team can interpret and act on without the owner.” Two constructs live in that clause. The first is sensemaking — Weick’s [10] account of the routinely underestimated work of turning data into meaning. A metric does not announce what it means or what to do about it; someone has to read it, frame it against expectation, and decide. The second is the distribution of that work beyond a single person. At Levels 1 through 4, the interpretive act can still sit with the owner; at Level 5, the organization itself can perform it. This is the property the broader maturity model reserves for its highest band, the closed loop, in which observation triggers documented action whose outcome is measured and fed back. The data-driven-decision-making literature is where this stops being a matter of taste and becomes a matter of measured outcome: Brynjolfsson, Hitt, and Kim [3], using instrumental-variable methods to address reverse causality, found that firms which genuinely run on data and analytics show output and productivity roughly five to six percent higher than their technology and other investments would predict, with parallel effects on asset utilization, return on equity, and market value. In the marketing domain specifically, Germann, Lilien, and Rangaswamy [4] found that greater deployment of marketing analytics is associated with higher return on assets. The Level-5 firm is the firm positioned to capture that premium, because the capability the premium rewards is not the possession of analytics but their use in decisions, distributed across a team that can act on them.

5. What the firm-performance evidence actually shows and what it does not

The convergent finding across independent literatures is directionally consistent: the disciplined use of measurement to inform decisions is associated with better firm performance [1][3][4][5]. Three honest qualifications are required, and a brief that omitted them would be a sales document.

• The effect is firmer for decision-making in general than for measurement systems in particular. Data-driven decision-making shows a robust association with performance [3]. The marketing-specific evidence is genuinely mixed: Homburg, Artz, and Wieseke [2] found that comprehensive marketing performance measurement systems do not directly improve firm performance, arguing the effect is indirect — operating through the quality of the decisions the system enables, not through measurement itself. This is not a contradiction of the ladder; it is a precise statement of why the top rung matters. Measurement that is not metabolized into better decisions does not pay. The mechanism is the action, not the dashboard.
• Causality is hard. Even the strongest study here [3] had to deploy instrumental variables precisely because firms that adopt data-driven decision-making differ in other ways from those that do not. The association is robust; “install metrics and performance will follow” overstates it.
• The construct is partly self-reported, with all the bias §2 names. A firm proud of its dashboards will tend to rate its own Insight maturity generously.

The defensible claim is therefore narrower and stronger than a slogan: firms that convert measurement into distributed, validated, acted-upon decisions tend to outperform; the rung that does this work is the top one; and most firms never reach it because the binding constraint has never been the data, it has been the discipline of choosing the right few measures and the organizational capacity to act on them without the owner.

6. What artificial intelligence does to the answer

The argument proceeds in three moves.

Move 1 — AI commoditizes the lower rungs.

Everything that separates Levels 1 through 3 from Level 4, capturing metrics, decomposing revenue into attributable signals, generating dashboards, and surfacing anomalies, is precisely the work that generative and analytic AI now performs cheaply and on demand. The historical barrier to leaving Level 1 or 2 was cost and skill: building attribution and a dashboard took an analyst the firm did not have. That barrier is falling. The consequence is not that measurement matters less; it is that being stuck at the bottom of the ladder is no longer explained by scarcity. When producing the metric is nearly free, remaining at “a feeling” or “revenue only” is no longer a resource constraint. It is a discipline failure and an increasingly visible one.

Move 2 — AI raises the premium on the top rungs and relocates the scarce skill.

If generating measures is now abundant, the scarce capabilities are the two the top of the ladder names: choosing the few that matter, and acting on them. AI makes the three-hundred-metric dashboard [5] not a cautionary tale but the default output — it will happily generate every measure that can be computed. The discipline of the short, valid, decision-relevant list becomes more valuable as measurement becomes more abundant, not less, for the same reason editing becomes more valuable as writing becomes cheaper. And the interpretation-and-action layer, Weick’s sensemaking [10], distributed across an answerable team,  is precisely the layer AI cannot occupy on its own. A model can produce the read-out; it cannot be held responsible for the decision the read-out implies. This is the same structural point the first brief in this series made about named accountability, arriving here from the measurement side: when a system can act, its errors become actions [13], and the value of a human who can interpret, judge, and answer for the result rises rather than falls.

Move 3 — the current evidence already shows the split.

The 2026 enterprise data, vendor and consultancy-reported, and flagged as such [12][13] ,  describes a population sorting into two groups along almost exactly the Insight ladder. Adoption is near-universal: by McKinsey’s account, a large majority of organizations now use AI in at least one function. But the realized value is concentrated in a small set of high performers, and what distinguishes them is not access to the technology, which is commoditized, but operating discipline. High performers are far more likely to set outcome-based objectives tied to business metrics, to rigorously measure adoption, quality, and results, and to maintain defined human-in-the-loop validation of model output; by one reported figure, sixty-five percent of high performers versus twenty-three percent of the rest [12]. Translated into the language of this instrument, the firms capturing value from AI are the firms already operating at Levels 4 and 5 of the Insight dimension. AI did not relax the requirement to know whether your growth engine is working; it raised the stakes of the answer and accelerated the separation between firms that can answer and firms that cannot.

The thesis, stated plainly. Cheap, abundant AI commoditizes the production of measurement and pushes the bottom of this ladder out from under firms that were resting there on the excuse of cost. It simultaneously raises the premium on the two capabilities at the top — disciplined metric selection and distributed, answerable action — that it cannot supply itself. The Insight dimension does not depreciate in the AI era. It becomes the threshold test for whether a firm can use AI at all.

7. Where this item sits: Insight as the loop-closing dimension

A construct-validity note, in the spirit of the series. Across the four briefs, the Growth pillar decomposes into Team (who owns growth), Process (where the method lives), Tools (where the record lives), and Insight (whether it is working, and what is done about it). The first three are capabilities a firm can build and then possess. Insight is different in kind: it is the feedback mechanism that makes the other three self-correcting. A named growth owner, a codified playbook, and a clean system of record are necessary, but without the Insight loop they form an open-loop system, well-built, and blind to its own results. This is a point in the instrument’s favor: by separating Insight from the other three dimensions, the OQ Score isolates the construct the maturity literature identifies as the actual difference between a Refined and a Mastered organization, not more capability, but capability that observes and corrects itself.

It also carries a measurement risk worth flagging in the same breath. The Level-5 anchor of this item — “interpret and act on … without the owner,” a closed loop — uses the same language the overall scale reserves for its top band, Mastered (“observations trigger documented improvements with measured outcomes”). The instrument should ensure that the Growth · Insight · Level-5 item and the global Mastered definition are measuring distinct things,  the former, the loop within the growth function; the latter, the loop as a property of the whole business,  rather than scoring the same closed-loop construct twice and inflating the correlation between them. The distinction is defensible and, in the author’s judgment, real; it should be made explicit in the methodology so a numerate evaluator does not mistake it for double-counting.

8. What this evidence does and does not establish

What the peer-reviewed record supports. That intuition about frequency and importance is systematically biased toward the salient and recent [6]; that disciplined, data-driven decision-making is associated with measurably higher productivity and firm performance [3], including in the marketing domain [4]; that firms which validate the link between their measures and outcomes outperform those that do not [5]; and that measurement ability is associated with both firm performance and executive confidence in marketing [1].

What the evidence does not support. It does not establish that installing metrics causes performance. The strongest study isolates the effect precisely because adopters differ in other ways [3], and the marketing-specific evidence is mixed, with at least one major study finding no direct effect and an indirect one operating through decision quality [2]. It does not establish the magnitudes of the 2026 AI figures, which are vendor- or consultancy-reported, vary by source and definition, and should be read as directional planning inputs [12][13]. And it does not establish that the maturity levels are equal-interval; they are ordered states, and averaging them implies more precision than the scale possesses.

On the instrument. The OQ Score’s Insight item is face-valid and, in the author’s judgment, well-aligned to the construct the literature identifies as operative, validated, parsimonious, and acted-upon measurement, not metric count. It has not been validated for predictive power against external performance outcomes, and the instrument’s own methodology says so. This brief evaluates the constructs the item measures; it does not certify the item’s predictive validity, which would require empirical work the instrument has not yet done.

Conflict of interest. This document is produced by the organization that produces the OQ Score. That is a material interest, and the analysis of the instrument should be weighed accordingly.

For all evidence based research and academic papers, visit this link for updates. https://boisusa.com/category/inside-the-oq-score/ 

References

1. O’Sullivan, D., & Abela, A. V. (2007). Marketing performance measurement ability and firm performance. Journal of Marketing, 71(2), 79–93. https://doi.org/10.1509/jmkg.71.2.079
2. Homburg, C., Artz, M., & Wieseke, J. (2012). Marketing performance measurement systems: Does comprehensiveness really improve performance? Journal of Marketing, 76(3), 56–77. https://doi.org/10.1509/jm.09.0487
3. Brynjolfsson, E., Hitt, L. M., & Kim, H. H. (2011). Strength in numbers: How does data-driven decisionmaking affect firm performance? ICIS 2011 Proceedings / SSRN. https://papers.ssrn.com/sol3/papers.cfm?abstract_id=1819486
4. Germann, F., Lilien, G. L., & Rangaswamy, A. (2013). Performance implications of deploying marketing analytics. International Journal of Research in Marketing, 30(2), 114–128. https://doi.org/10.1016/j.ijresmar.2012.10.001
5. Ittner, C. D., & Larcker, D. F. (2003). Coming up short on nonfinancial performance measurement. Harvard Business Review, 81(11), 88–95. https://hbr.org/2003/11/coming-up-short-on-nonfinancial-performance-measurement
6. Tversky, A., & Kahneman, D. (1974). Judgment under uncertainty: Heuristics and biases. Science, 185(4157), 1124–1131. https://doi.org/10.1126/science.185.4157.1124 — Availability heuristic: Tversky, A., & Kahneman, D. (1973). Availability: A heuristic for judging frequency and probability. Cognitive Psychology, 5(2), 207–232. https://doi.org/10.1016/0010-0285(73)90033-9
7. Pfeffer, J., & Sutton, R. I. (2006). Evidence-based management. Harvard Business Review, 84(1), 62–74. https://hbr.org/2006/01/evidence-based-management — Book: Hard Facts, Dangerous Half-Truths, and Total Nonsense. HBS Press.
8. Kaplan, R. S., & Norton, D. P. (1992). The balanced scorecard—Measures that drive performance. Harvard Business Review, 70(1), 71–79. https://hbr.org/1992/01/the-balanced-scorecard-measures-that-drive-performance-2
9. Simons, R. (1995). Levers of Control: How Managers Use Innovative Control Systems to Drive Strategic Renewal. Harvard Business School Press. https://www.hbs.edu/faculty/Pages/item.aspx?num=257
10. Weick, K. E. (1995). Sensemaking in Organizations. SAGE Publications. https://www.sagepub.com/shop/buy-a-book/sensemaking-in-organizations-1-4988
11. Davenport, T. H., & Harris, J. G. (2007). Competing on Analytics: The New Science of Winning. Harvard Business Review Press. https://store.hbr.org/product/competing-on-analytics-updated-with-a-new-introduction-the-new-science-of-winning/10157
12. McKinsey & Company. (November 2025). The State of AI in 2025: Agents, Innovation, and Transformation. https://www.mckinsey.com/capabilities/quantumblack/our-insights/the-state-of-ai — Vendor/consultancy-reported; figures directional.
13. McKinsey & Company. (March 2026). State of AI Trust in 2026: Shifting to the Agentic Era. https://www.mckinsey.com/capabilities/tech-and-ai/our-insights/tech-forward/state-of-ai-trust-in-2026-shifting-to-the-agentic-era — Vendor/consultancy-reported; figures directional.
14. Ries, E. (2011). The Lean Startup. Crown Business. https://www.penguinrandomhouse.com/books/210088/the-lean-startup-by-eric-ries/ (Practitioner source; cited for the vanity-vs-actionable-metric distinction.)