Why the future of retail isn't data, but automated decisions

Ten years ago, modern retail's mantra emerged almost without debate: "the future belongs to data-driven retailers." That formula, appealing in its simplicity, structured a decade of investment, hiring, and strategic narrative. Every player in the sector, from global giants to national chains, adopted the same compass: collect more data, store it better, model it better, visualize it better. Data had become retail competition's playground.

Ten years later, the observation is clear: that compass pointed in the right general direction, but it caused the real target to be missed. Retailers who massively invested in data find themselves today with exceptional infrastructure — petabytes of collected data, trained ML models, highly qualified data teams — and operational performance that has, on average, barely progressed. Meanwhile, the retailers who structurally widened their performance gap — Inditex, Costco, Uniqlo, certain emerging e-commerce players — didn't win because they had more data. They won because they knew how to turn it into executed decisions, at a speed and scale their competitors can't replicate.

This observation, increasingly impossible to ignore, signals a paradigm shift. The future of retail won't be defined by data quantity or quality — those dimensions are now largely commoditized. It will be defined by the ability to automate decision-making — meaning to industrialize, at scale, the operational arbitration that turns data into performance. And that transition, barely starting, will deeply redraw the sector's competitive landscape over the next five to ten years.

This article looks at what's coming, and why this transition is both inevitable and radical. Not an incremental improvement of current practices. A change in the very nature of what running a retailer means.

The end of a cycle: why data is no longer a competitive edge

To understand what's coming, you have to accept a truth few vendors and consultants want to state clearly: data, as an asset, has become commoditized. This commoditization is recent, but massive and irreversible.

Ten years ago, accessing modern data infrastructure was a real competitive edge. Building a data warehouse cost millions, mobilized rare specialized teams, took years. Having a performant BI platform clearly distinguished leaders from followers. Having internal data scientists was a luxury few chains could afford.

Today, these capabilities are accessible to any retailer with a reasonable budget. Snowflake, Databricks, BigQuery, Power BI, Tableau, Looker: data platforms are mature, documented, deployable in months. Data skills, still rare five years ago, are now widely available on the market. ML models, once proprietary, are open source, pre-trained, accessible via API. Data is no longer a differentiation ground — it has become a prerequisite everyone masters at roughly the same level.

This commoditization has a major strategic consequence: the marginal return on data investments keeps degrading. Adding a new dashboard, refining a model, enriching a data lake produces today much less differentiation than five years ago. Not because these investments don't create value anymore — they still do — but because your competitors are doing exactly the same, with the same tools, reaching the same results. The field has leveled.

Retail competition has shifted. It no longer plays out upstream (who has the best data?) but downstream (who knows how to exploit it most effectively in operations?). And it's downstream that the next differentiation cycle will play out.

The new ground: decision automation

If data is now commoditized, what isn't yet? The answer is clear: the ability to turn data into executed operational decisions, at scale, continuously, with high quality. It's in this zone, structurally under-invested during the previous decade, that the next competitive edge plays out.

To understand what "automating decision" means, you have to distinguish three levels of industrialization.

Level 1: industrialize data collection and visualization. That's what retailers have done over the past ten years. Largely solved today — and therefore no longer differentiating.

Level 2: industrialize the formulation of recommendations. That's what some are starting to do with their predictive models and analytical platforms. Still developing, but already accessible. Not enough on its own, because an unexecuted recommendation creates no value.

Level 3: industrialize the full decision → execution → learning chain. That's the playground that's opening. For now, very few retailers truly operate there. It's the one that will define the next decade's leaders.

This third level isn't an extension of the previous one. It's a qualitative leap that requires different architecture, organization, and culture. A platform built to produce executable decisions. Formalized, executable business rules. Execution without breakage between decision and floor. A learning loop that continuously adjusts. And — perhaps more important than the rest — an organization that accepts that routine operational decision is no longer an exclusive human prerogative, but an industrialized function piloted by the teams.

Why automating decision changes everything

When a retailer crosses over to automated decision, what changes isn't marginal. It's a shift in operational physics.

First shift: speed. Where everyday human decisions typically take five to ten days between signal and action, automated decisions take hours, sometimes minutes. At a season's scale, across tens of thousands of SKU/store pairs, this cumulated speed difference represents several complete cycles of operational learning. The automated retailer learns ten times faster than its non-automated competitor — and in two or three seasons, this learning gap becomes structurally insurmountable.

Second shift: scale. A human team, however competent, can seriously handle a few hundred decisions per week across an extended network. An automated system can handle tens of thousands, without quality degradation. This scale difference isn't a marginal factor — it's a multiplicative factor that changes the nature of steering. The automated retailer steers every SKU in every store with a level of care the manual retailer can only give to its strategic products.

Third shift: consistency. When decisions are made by scattered human teams, they inevitably drift — one merchandiser applies a rule slightly differently from another, one store interprets an instruction differently from its neighbor, one season imperfectly reuses the lessons of the previous one. When decisions are automated, these drifts disappear by construction. Consistency is no longer a target to aim for; it's a default property of the system.

Fourth shift: learning capitalization. In a human organization, learning is fragile. It depends on individual memory, turnover, informal transmission. In an automated system, every decision and its result is recorded, analyzed, integrated into the next calibration. Operational knowledge becomes an accumulated asset that grows season after season — and doesn't evaporate when someone changes role or leaves.

The sum of these four effects produces a performance difference that isn't marginal. Retailers who've crossed over typically report several points of net margin recovered — not from luck or exceptional effort, but as a mechanical consequence of industrializing decision.

The classic objections (and why they all fold)

When you raise automated decision in retail, you systematically hear the same objections. They deserve to be taken seriously — but they all fold, one by one, under honest examination.

Objection 1: "You can't automate a retail decision, it's too contextual." The argument is intuitive but empirically false. Modern retail has already let itself be automated on dimensions once deemed irreducibly human — product recommendation, price personalization, ad allocation. Each time, the "too contextual" argument preceded an automation wave that transformed the craft. Routine operational decision (markdown, transfer, replenishment) is on exactly the same trajectory.

Objection 2: "Our teams won't accept it." This objection masks a confusion. Automating decision doesn't aim to replace teams — it aims to free them from repetitive mechanics so they can focus on high-value-added trade-offs. Retailers who've crossed over almost always observe the opposite of expected resistance: operations teams, freed from dozens of hours of weekly Excel processing, rediscover the interest of their craft. Resistance, when it exists, almost always comes from a badly designed rollout — not from the principle itself.

Objection 3: "We'll lose our know-how." Exactly the opposite happens. When know-how stays in people's heads, it's fragile, untransmissible, vulnerable to turnover. When it's formalized in a system, it becomes a durable company asset, enriching with each cycle rather than eroding. Automation isn't loss of know-how — it's its capitalization.

Objection 4: "Models make mistakes humans wouldn't." True in some cases, but the argument is asymmetric. Humans also make mistakes — far more, actually, because they don't have the bandwidth to handle every case seriously. The right question isn't "do models make mistakes?" but "is the global error rate of the system (humans + machines) higher or lower than the error rate of humans alone?". In almost every studied case, the answer is unambiguously in favor of the well-designed hybrid system.

Objection 5: "Our business is too specific." The cultural-exception argument, heard in every sector before automation. It's almost always true at the margin — every retailer does have specificities — and almost always false at the core. The principles of retail operational decision are universal (arbitrating between stock, velocity and margin under constraints), even if their implementation varies. Modern platforms are built to absorb that variability, not ignore it.

These five objections all fold — but they slow adoption. And it's precisely that slowdown that creates an opportunity window for retailers who dare to cross over earlier than their competitors.

The widening gap: the next five-year window

Sector transition history offers a useful frame. When a sector shifts to a new paradigm — 19th-century industrialization, 1980s computerization, 2000s e-commerce — the gap between early adopters and laggards widens non-linearly. The first take a lead that becomes progressively insurmountable, because it compounds over time: better calibration, more learning data, more seasoned teams, more mature processes.

The transition to automated decision follows the same logic. Retailers engaging on this ground today accumulate, season after season, operational experience and calibration quality their competitors won't be able to catch up with quickly — even by investing massively later. At a three-to-five-year horizon, the gap will be structural.

This dynamic has a major strategic consequence for retail leadership today. Postponing the transition to automated decision isn't a neutral choice. It's not "wait to decide better" — it's "accept that others will take a lead we won't catch up with." The opportunity cost of waiting grows, and at some point becomes determining for the chain's competitive position.

Conversely, retailers who dare to engage now, even imperfectly, are the ones who'll define the next decade's competitive landscape. Not necessarily because they'll have the best platforms at the start — but because they'll have started accumulating, earlier than others, the operational experience that turns promise into performance.

What this changes for retail leadership today

For retailer executives watching this transition, three concrete implications deserve emphasis.

First implication: reposition data investments. Not reduce data budgets — data remains essential as raw material. But redirect marginal investment from upstream (collection, storage, BI) to downstream (decision, execution, learning). This redirection is cultural as much as budgetary. It requires accepting that the next value wave won't come from the layers you know well — but from a layer you have to learn to build.

Second implication: reposition human roles. If routine operational decision is automated, what do teams do? The answer isn't "less" — it's "something else." Assortment strategy, supplier negotiations, customer experience, experimentation, structural arbitrations: all functions where human value-add remains irreplaceable and where teams can redeploy once freed from repetitive mechanics. This transition isn't disguised layoffs — it's a move up the value chain.

Third implication: accept transformation as a transformation project, not a tech project. Decision automation isn't a tool you install. It's a deep transformation of how the chain steers its operations, involving data architecture, business processes, team roles, management culture. Retailers who'll succeed in this transition are the ones who'll run it as a full transformation project, not as buying another piece of software.

The Solya approach: automated decision as architectural principle

That's precisely Solya's mission. Not yet another data science tool. Not yet another sophisticated dashboard. A decision and execution platform built, from the start, to carry the industrialization of retail decision at network scale — meaning to address exactly the differentiation ground that's opening.

Concretely, Solya connects to your existing data sources — POS, ERP, e-commerce, supply chain — and rebuilds a live view of the network at the SKU/store level. The decision engine continuously scans that view to identify action opportunities, formulates contextualized recommendations embedding your business rules and operational constraints, and propagates validated decisions to your execution systems without breakage. The observed effect feeds the learning loop that refines subsequent decisions. Your teams keep the hand on every structural trade-off — they define strategy, validate sensitive cases, adjust parameters. Solya handles operational mechanics: the thousands of routine decisions no one can manually handle at the required cadence and quality.

The result goes beyond immediate gains (overstock reduction, lower markdowns, improved service rate). It lays the foundations for a longer-term transformation: an organization that learns from every decision, capitalizes its know-how in a system rather than in individual heads, and takes operational lead season after season over non-equipped competitors. That's exactly the kind of compounded competitive advantage that distinguishes, at a three-to-five-year horizon, retailers who got ahead of the automated decision transition from those who'll have endured it.

The real question to ask

In three to five years, when automated decision has become the new retail standard the way data became one in the past decade, where will your chain be positioned? Among those who took the early lead and accumulated the operational experience that flows from it? Or among those scrambling to close, in panic, a gap that has become hard to bridge?

This question isn't rhetorical. For the first time since e-commerce's arrival, it is retail's structuring strategic question. Data was a differentiation ground between 2015 and 2025. Automated decision will be one between 2025 and 2035. Retailers who've understood this transition and engaged it early will define the next decade's competitive landscape. The others will adapt to a world they didn't help shape.

The future of retail isn't data. Data was the previous step, now largely commoditized. The future of retail is automated decision — meaning the ability, still rare today, to turn the latent intelligence of data into operational performance executed at scale, continuously, with a quality manual methods can no longer reach.

That's the transition redefining the sector. And it, more than any other transformation, will determine who leads and who follows in retail over the next decade.

---

Where are you on the transition to automated decisions?

At Solya, we offer retail leadership teams a personalized 30-minute diagnostic to assess, on your own context, your current positioning on this transition — and identify the first high-leverage use cases to engage the move before the competitive gap becomes hard to close.

👉 [Book your Solya diagnostic] — 30 minutes, by video, with one of our retail experts.

You'll walk away with:

• An evaluation of your current maturity on the decision-execution chain
• An estimate of the value potential accessible by industrializing your operational decisions
• The first priority use cases to engage the transition at the right pace and risk level