From data ness to AI-driven growth: The 5-year journey of RGM transformation
A practical roadmap to scale RGM: data foundations, adoption, and retailer-ready decision making.
In the first half of the year, a series of conversations across the CPG ecosystem sparked a simple realization: there are too many valuable RGM lessons that surface in executive rooms—and then disappear. This article captures the insights that came up repeatedly in discussions with CPG leaders, technology partners, and revenue management experts.
The message was consistent: it’s not about the AI, it’s about the foundation.
This is the first contribution in the series, don’t miss the follow-up on where AI really adds value in RGM (and where it doesn’t).
Key Takeaways:
Data quality is a bottleneck: without trusted inputs, RGM decisions and models break.
Technology fails without adoption: process change and new ways of working must follow
Explainability builds retailer credibility: teams need to clearly justify the “why” behind action
The data problem nobody wants to admit
When leaders in advanced analytics join established CPG companies, they discover something surprising: the sophisticated machine learning models already exist, but the data itself is a mess.
This is the first and most important lesson of RGM transformation. Most organizations spend 80% of their time gathering data, cleaning it, and formatting it into something analyzable. Only a fraction of that effort goes to actual insights and decision-making. The rest disappears into meetings, debates, and implementation challenges.
The problem compounds over time. Early in an RGM journey, companies build quick wins with whatever data is available. They create dashboards and simulation capabilities. But these initial solutions, while useful, often mask a deeper problem: the underlying data is unreliable.
After several years of building increasingly sophisticated analytical layers on top of imperfect data, organizations hit what technologists call “the law of technical debt.” Everything you’ve built starts to crumble because the foundation was never solid.
This is where most companies plateau. They’ve invested in tools. They’ve trained people. But they can’t scale because the data quality remains the real bottleneck.
The implication is profound: before you implement AI, you need to solve the data problem. This is foundational.
> A line that stood out
Everything you build on top of messy data is garbage. Rubbish in, rubbish out. Because the AI piece doesn’t need a report or anything. It reads immediately from the data and does everything from there.
The organizational reality: why technology alone fails
Even with perfect data and sophisticated technology, something else often goes wrong: people don’t use it because organizations underestimate how much change is actually required.
When new tools are introduced, they typically demand that people work differently. Sales teams need to think about trade terms and promotional ROI, not just volume. Marketing needs to understand elasticity and baseline sales. Finance needs to engage earlier in promotional planning.
The adoption challenge is amplified by structural realities. In most CPG organizations, RGM practitioners come from diverse backgrounds (finance, marketing, or sales). Diversity is value, but it also means people are learning RGM itself while simultaneously learning new tools and new ways of working.
And they’re doing this in an environment where staff turnover is high. By the time someone truly understands the system, they’ve often moved to a different role.
Add to this the fact that organizational processes often haven’t evolved to support the new tools. An RGM team might have a sophisticated optimization platform, but if the questions being asked remain the same—if success is still measured by last year’s metrics—then the tool sits unused.
The turning point for organizations that succeed is when leadership stops focusing on the technology and starts focusing on the people using it.
The change management imperative: the human side
Technology alone isn’t enough. The change management challenge is equally important—and it’s where many organizations stumble.
One framework that emerged across multiple discussions was the ADKAR model: Awareness, Desire, Knowledge, Ability, and Reinforcement.
The critical insight was where most organizations go wrong: they skip straight to the “Knowledge” phase.
What they’ve missed is the foundational work of helping people understand why change is necessary, what’s in it for them personally, and what the risks are if the organization doesn’t change.
This communication needs to happen before deployment. It needs to be continuous, not a one-time event.
> A line that stood out
You need to start with continuous communication: most organizations talk about what and how but skip the why. And that’s where adoption fails.
The organizational design question
An emerging theme across discussions challenges a fundamental assumption: should RGM be a dedicated function at all?
One perspective suggests that if technology could reduce the complexity of RGM decisions—by limiting the number of choices presented and explaining the reasoning behind each recommendation—you might not need people with such diverse skill sets.
This is speculative, but it points to an important truth: the organizational structure of RGM will likely evolve as technology matures.
The retailer dynamics: the constraint nobody controls
There’s a constraint that often gets overlooked in RGM discussions: a perfectly optimized promotional plan means nothing if the retailer says no.
Retailers are often risk-averse. They prefer incremental changes to radical shifts.
When you come to them with a recommendation, you need to be able to explain it in their terms, not in the language of algorithms and optimization.
This is why storytelling matters.
> A line that stood out
Ultimately you need to sell it into a retailer who might ask you why you’re proposing this plan. If you tell them ‘because the computer says so,’ you’ve lost credibility.
The five-step pattern: what actually works
Based on the experiences shared across discussions, a pattern emerges about how realistic RGM transformation unfolds.
The first phase typically focuses on data foundation-building…
The second phase involves integration…
The third phase is where sophistication increases…
The fourth and fifth phases involve genuine transformation…
This five-step timeline isn’t arbitrary. It reflects the time investment required to move through each phase with sufficient rigor that the transformation sticks.
> A line that stood out
We started the journey where we weren’t doing any promo optimization. We did quick wins to prove value, then moved to web-based solutions, then added PPA modeling. But the real transformation only happened when we stopped focusing on the tool and started focusing on the people and processes around it.
Conclusion: from mess to impact
The journey from data mess to AI-driven growth isn’t quick or easy. But it’s necessary.
The technology is ready. The platforms exist. The methodologies have been tested.
What remains is the hard work of organizational transformation: getting data right, preparing people for change, redesigning processes, and building the discipline to embed new ways of working into daily routines.
> A line that stood out
Technology is only 10% of the equation. It’s really the change management, the people, and the processes that make the difference.
Closing thoughts
Foundation comes first—always.
Data quality, organizational readiness, and process design determine whether RGM scales or stalls.
A realistic roadmap takes time (often around five years) because adoption, operating model, and ways of working must change with the tools.
When teams can explain recommendations in retailer terms, RGM becomes repeatable, credible, and resilien, not fragile.