Evidence
Do data-driven companies perform better?
Yes, there is an advantage — but it shows up when data changes decisions, not when a company buys tools. What 15 years of studies and the 2026 surveys say.
Anyone who argues for evidence-based decisions has an annoying obligation: to apply the same rule to themselves. If the question this magazine keeps asking is “how do you know?”, it applies to Data-Driven Culture itself.
So, do companies that decide with data perform better?
The short answer is yes, with one condition that changes everything: the advantage appears when data changes decisions, not when a company piles up systems. The 2026 figures on artificial intelligence make that condition even more visible.
The first clue: productivity
In 2011, Erik Brynjolfsson, Lorin Hitt and Heekyung Kim analyzed 179 large publicly traded companies. Those that adopted data-driven decision-making had output and productivity 5% to 6% higher than expected, even after accounting for their other technology investments.
The key detail is “even after accounting for.” The difference was not about who bought more software, but about who brought data into the decision.
Eleven years later, Thuy Duong Oesterreich, Eduard Anton and Frank Teuteberg pooled 125 studies from 26 countries in a meta-analysis. They found an association between analytics and better performance — and a greater weight of human and organizational factors, such as skills, management and culture, than of purely technical ones.
2026: AI reached people before it reached results
McKinsey’s annual AI survey, published in August 2026 with 1,719 participants, shows a gap any manager will recognize.
Source: McKinsey, The State of AI (Aug. 2026). Online survey, May–June 2026, self-reported answers. The three bars answer different questions: they are not stages of a funnel.
Eight in ten people say AI has improved their own work. Just over a third see an effect on operating profit. And only 6% make it into the group McKinsey calls high performers.
You cannot subtract one number from the other and call the difference “lost productivity.” But the contrast shows where the work is: turning individual gains into company results requires redesigning processes, integrating data and deciding where the tool is worth it.
Size matters too. Among companies with revenue above US$1 billion, the share already scaling AI agents in at least one function rose from 27% to 40% in a year. Among smaller ones, it stayed flat at 22%.
Source: McKinsey, The State of AI (Aug. 2026). Different samples each year.
Those already reaping results grow faster
BCG’s applied AI index, from September 2026, surveyed 1,330 executives. The consultancy classifies 7.5% of companies as future-built — the most mature in their use of AI — and another 41% as companies already scaling the technology and generating some value from it.
Compared with laggards, the most mature companies show much better numbers on three financial measures.
Source: BCG, press release of Sept. 30, 2026 and The Formula for Agentic AI Value. Groups defined by BCG itself; a comparison, not an experiment. Scale from 0× to 3×.
It is a strong association, but it does not prove that AI alone caused the difference. Companies that were already more efficient can invest more in AI — and reap more because of it. McKinsey’s 6% and BCG’s 7.5% are not the same group either: each consultancy uses its own criteria.
There is a clue about what separates them. In April 2026, Gartner reported that organizations with successful AI initiatives invest up to four times more, as a share of revenue, in foundations such as data quality, governance, workforce readiness and change management.
And in Brazil?
Among Brazilian small businesses, digitalization has advanced fast. Sebrae’s Digital Maturity Index rose from 35 to 39 points, on a 0–80 scale, between 2024 and 2026.
Source: Agência Sebrae, Oct. 23, 2026 (in Portuguese). In 2026, most businesses (60%) are in the intermediate “emerging” group. Bar scale: 0% to 100%.
The beginners group has practically disappeared. But 60% are halfway there and only 9% are digital leaders. And digital maturity is not the same as deciding better: the index measures digital tools and practices, not margins or profit. To know whether this progress turned into results, you would need to follow the same companies and their numbers over time.
The mechanism: the decision in the middle
A study published in 2026 in the Journal of Business Research by Xenia Mamakou helps put the pieces together. Analyzing professionals who use analytics at work, it concludes that success does not come from isolated investments but from combinations of capabilities: quality data, tools, skills and business knowledge. And it points to decision quality as the link that turns those capabilities into performance — with company size changing which combination works.
It is the same conclusion that runs through the studies since 2011, now with a more precise name: good data and available tools are the foundation; value appears when someone decides differently because of them.
How to test this in your company
The studies compare companies with one another. In yours, the question is different: which decisions changed because of data — and how much did that yield? A simple roadmap:
- Pick one concrete decision. How much to buy, what price to charge, whom to serve first.
- Record the “before.” How the decision was made and what the result was, for a few weeks.
- Change one thing and document it. What is now used in the decision, who decides, when.
- Compare carefully. Account for seasonality and other changes in the period. If you can, compare with a similar store, team or period that did not change.
- Do the full math. The gain attributable to the change minus the cost of implementing and maintaining it.
Without an honest point of comparison, the result is a management estimate — useful, but not proof.
So?
The accumulated evidence favors companies that decide with data. But it does not favor buying technology indiscriminately. In 2011, the advantage showed up in productivity. In 2022, the synthesis of studies pointed to people, management and culture. In 2026, AI repeats the lesson: adoption is easy to measure; results only come when the decision changes.
It would make no sense to end with absolute certainty. So we end with the question this magazine always asks: in your company, what changed after the data arrived?
Sources: Brynjolfsson, Hitt & Kim (2011), ICIS · Oesterreich, Anton & Teuteberg (2022), Information & Management · McKinsey, The State of AI (Aug. 2026) · BCG, press release, Sept. 30, 2026 · BCG, The Formula for Agentic AI Value (2026) · Gartner, Apr. 16, 2026 · Agência Sebrae, Oct. 23, 2026 · Mamakou (2026), Journal of Business Research.
Method note: only figures confirmed in the original publications or official releases were included. Self-reported surveys and comparisons between groups show association, not cause. This story replaces the version published on September 25, 2026.