The struggle to be data-driven
The feature factory problem
For decades, organizations have chased the goal of being “data-driven.” Yet despite huge investments and leaps in technology, how many really are? Too often, decisions are still based on gut instinct or metrics chosen to confirm a preferred story. In that environment, releasing features becomes the default signal of progress: Teams are rewarded for speed and delivery, while the harder question—”Did any of this create value?”—goes unanswered.
The data is sobering. Across large product orgs, only about 10–30% of ideas end up delivering a measurable business impact when tested. The rest are neutral or negative. Most shipped features add cost, complexity, and ongoing maintenance without moving the outcomes that matter, creating a hidden tax on future execution and usability.
As product thinkers like John Cutler and Itamer Gilad have noted, this pattern reflects a “feature factory” culture, where output becomes the measure of success, even when it isn’t tied to real impact. Teams caught in this cycle mistake activity for progress, shipping features faster but learning slower.
To escape, organizations must shift their mindset from a feature factory to a value generator: one that defines progress by outcomes achieved, not by the volume of launches.
However, escaping the feature factory isn’t just a mindset shift; it’s a systems problem. Despite years of effort to move from outputs to outcomes, most organizations are still trapped by the deeper structural forces that disconnect work from impact.
The “messy middle” is the root cause of the struggle
Every company ultimately wants to move its business KPIs: the long-term health measures like revenue, customer growth, and retention that show whether the business is thriving. These numbers are essential, but they share a common limitation: they are lag measures. As the authors of The 4 Disciplines of Execution put it, lag measures are like checking the scoreboard after the game is already over. They tell you the outcome, only after it’s too late to do anything.
Meanwhile, day-to-day work happens at a different tempo. Teams ship features, run campaigns, and design experiments that may nudge local metrics such as clicks, signups, or conversion rates. Yet the connection between those local improvements and the company’s top-line KPIs is often unclear.
This disconnect creates what many organizations experience as the “messy middle”: a gap between everyday activity and the outcomes that matter most. Leaders are left uncertain about which bets truly drive growth, while teams feel powerless to influence the numbers that define success.
Executives & Leaders
Problems
Revenue is flattening, but the root cause is unclear
Dashboards show what happened, not why
Consequences
Feel blind and deflated
Uncertainty spreads as teams can’t see how to make a real difference
Product Managers
Problems
Flooded with feature requests
Lacking a data-driven framework for prioritization
No clear map of which metrics truly drive impact
Consequences
Feel stuck in opinion battles and politics
Confidence erodes as prioritization feels arbitrary
Teams
Problems
Can’t trace how their work connects to outcomes or customer value
Strategy and daily reality drift apart
Consequences
Feel a growing dissonance and loss of purpose
Morale drops as faith in leadership and direction fades
Even well disciplined “data-driven” teams aren’t immune.
Let’s look at an example. A marketplace I worked with once set an OKR to increase average services per seller. The team hit its target—services per seller rose 39%—and celebrated. But soon after, seller retention fell 6% after low-engagement services flooded the marketplace.
By optimizing a local metric in isolation, they incentivized behavior that actually hurt the business. They had the data, the dashboards, and the OKRs—but without a model connecting initiatives to business value, they mistook movement for progress. Consequently, they invested millions in an initiative that actually hurt their business.
Initiative
Grow services
per seller
$
2M
per year
Metric
Average services per seller
39
%
Result
Seller retention rate
6
%
