A look at why digitizing Mexico’s 1.1M corner stores keeps failing — and what a neutral, capital-light platform could get right instead.
1. Introduction: An Old Problem, Seen From the Retailer Outward
Mexico’s traditional trade — its tienditas, abarrotes, changarros and corner stores — is usually described from the outside in: as a distribution problem to solve, a long tail to consolidate, or a market to digitize. This paper starts from the opposite direction. It asks what it would take to help the independent retailer at the center of that channel actually thrive — professionalizing, growing, and capturing more of the value their collective reach already creates — and treats everyone else (distributors, brand owners, financial institutions) as partners in that goal rather than the other way around.
The case for taking this channel seriously is not sentimental. As seen later in the document, it is the majority engine of how consumer goods reach Mexican households, at a scale that rivals or exceeds any single organized retail chain in the country. Building out the retailer’s world as it actually is today: how big and constrained this universe is, what it competes against, and how it is currently supplied, including a look at what has already been tried to change that, and why so much of it failed, the two approaches that are genuinely working right now, and the gap that persists even where they succeed. Taking all that into a set of design principles for serving this channel differently, and a recomended phased story — how the idea could move from a single pilot to a connected network, expanding in both geography and scope at a deliberate pace.
2. Sizing the Potential
Mexico’s traditional trade is not a niche. As of May 2026 the grocery and food retail trade counted 1,113,420 active economic units[1], the great majority of them micro-enterprises with ten or fewer employees. Independent industry data puts the traditional channel — small neighborhood stores, local markets and street vendors — at over one million points of sale, reaching nearly 100% of the population and employing around two million people directly, worth roughly 40% of the total value of Mexico’s mass-consumption industry[2].
The framing worth sitting with is this: this is not a fragmented long tail to be consolidated or replaced. Considered collectively, it is the majority engine of how consumer goods actually reach Mexican households — larger, by reach and by share of value, than any single organized retail chain in the country. The strategic question is not whether this network matters; every player discussed in this paper already agrees that it does. The question is who gets to help it professionalize, and on what terms.
3. The Mom-and-Pop Retailer: Constraints, Barriers and Untapped Potential
A typical independent store manages well over a hundred categories and more than a thousand SKUs, almost always without any dedicated inventory, purchasing or financial system beyond memory, a notebook, or a cash drawer. That is not a criticism of the retailer — it is close to structurally inevitable for a business run by one or two people, with the great majority of the country’s more than 1.1 million grocery and food retail units falling into the smallest, ten-or-fewer-employee size band[1]. Limited capital and limited time compound each other: a store owner who cannot afford to carry deep stock of everything, and who has no system to track what is moving, ends up simultaneously out of stock on fast-selling items and overstocked on slow-moving ones — tying up scarce capital on shelves in the same month sales are being lost elsewhere.
Set against this is a category of competitor that has solved exactly these problems at scale: professionally run convenience-store chains, foremost among them OXXO, whose owner FEMSA reports steady, growing revenue from that format[4]. These chains bring standardized store formats, centralized purchasing leverage, real-time inventory systems, professional merchandising, and a corporate balance sheet behind every location. None of that is available to a single independent owner acting alone. The asymmetry is real and, left alone, it widens: the same shopper base is being fought over by one side with systems, data and capital, and another side with none of the three.
Yet the independent channel is not simply a weaker version of the same business. It holds assets a chain cannot replicate at every corner: genuine proximity inside neighborhoods chains do not reach as densely, personal trust built over years with the same customers, informal credit relationships (“fiado”) that keep a loyal shopper coming back even in a tight week, and flexible hours suited to how the surrounding block actually lives. Individually, none of this is visible or valuable beyond a single block. Collectively, across more than a million storefronts reaching essentially the whole population, it is an enormous, almost entirely unorganized asset — in distribution reach, in advertising and communication surface, and in the granular, real-time market data that passes through those tills every day. The opportunity is not to make each store act more like a chain outlet; it is to let the network behave, collectively, like a competitor with systems and scale, without requiring anyone to own or run the stores themselves.
4. How Retailers Are Sourced Today
Supply reaches these stores through two broad routes. Direct distribution comes from the brand or producer itself, from an exclusive or multi-category distributor, or from an independent salesman who buys from a wholesaler and resells door to door. Indirect distribution means the retailer travels to a cash-and-carry wholesaler or price club to source product themselves — an added cost in time and transport for an owner who already has very little of either to spare.
Within direct distribution, two field models dominate. The pre-saler visits a store, takes the order, and leaves — the product itself arrives later on a separate delivery run, so the retailer is exposed to whatever happens between the visit and the delivery. The prompt-saler, or van-sales model, arrives with an already-stocked truck and sells and delivers in the same visit — faster, but the retailer can only buy what happens to be on that truck that day, not what they actually need.
What this generates
A field force that is expensive to run at consistent quality: supervising a large, geographically dispersed sales and delivery team well enough to guarantee service is inherently costly and hard to control, which is exactly the operational weight later attempts at digitization underestimated (Section 5), while offering limmited to what one person remembers (or has the time) to offer.
Fragmented visibility: any single distributor only ever sees the slice of a store’s business that runs through its own portfolio, so nobody — not the distributor, not the brand owner, not the retailer — has a full picture of that store’s actual demand across the hundred-plus categories it carries.
A cost-to-serve mismatch: a rep responsible for many small accounts has only a few minutes at each stop, so problems the retailer raises often never reach the people who could fix them, or arrive too aggregated across too many stores to act on.
Visit-frequency-driven stock-outs: because most routes run on a fixed weekly or biweekly cycle rather than on what a store has actually sold through, a fast-moving item can sit out of stock for days between visits while a slower one purchased to meet a minimum order sits unsold, tying up capital on both ends of the same problem.
A broken feedback loop: late deliveries, short orders or damaged product are typically reported to the same representative responsible for causing them, with no independent channel for that feedback to reach a distributor’s management in a form it can act on.
5. What’s Been Tried — and What Went Wrong
Digitizing this channel is not a new idea. Over the past several years, a wave of well-funded platforms across Latin America — and comparable ventures in other emerging markets — set out to solve the same problems described in Section 4 by taking over the entire stack at once: an ordering marketplace, owned logistics, and embedded retailer credit, in effect trying to replace the pre-saler/prompt-saler relationship rather than plug into it.
Frubana raised more than US$271 million and at its peak served roughly 80,000 restaurants and retailers across Mexico, Colombia and Brazil[5]. It exited Mexico and Colombia in February 2024 and wound down its last remaining market, Brazil, by July 2025[6]. Merqueo, a comparable Colombian player, closed around the same time[7]. The pattern is not unique to Latin America: Udaan, India’s largest B2B wholesale marketplace, saw its valuation cut by roughly 43.5% to about US$1.7 billion[8], and its own public retrospective concluded that the combination of marketplace, logistics and embedded credit became an unsustainable capital drain given how thin wholesale margins already are[9]. Wasoko and MaxAB, two of the largest B2B distribution platforms serving informal retail in Africa, were pushed into a distressed merger completed in August 2024 amid layoffs[10].
The common thread across all of these is not that digitizing traditional trade is a bad idea — it is that the specific combination of owning logistics, owning credit risk and owning the marketplace all at once asks a thin-margin business to also absorb a capital-intensive one, and inherits every risk of the existing distributor relationship (bad debts, service failures, underused warehouses) without the accumulated local knowledge that made that relationship work in the first place.
6. Two Models That Are Working — and Why
6.1 FEMSA’s “Juntos+”
Coca-Cola FEMSA officially launched Juntos+, a multicategory ordering app for tienditas, in Mexico and Brazil in March 2024[11]. It is already operating at a scale few startups ever reach: 1.1 million active monthly buyers, roughly US$2.5 billion in digital-channel revenue, 2 million customers served in Mexico, and just under two store visits a week on average[12]. It works because FEMSA is not starting from zero: it already runs one of the most capitalized, highest-frequency delivery networks touching small-format retail anywhere in Latin America, with decades of merchandising data and, crucially, deep and direct experience serving convenience retail through OXXO, the chain it owns and continues to expand[13]. Turning its own order-taking process into an app is a natural extension of infrastructure FEMSA already owns and operates — not a new capital bet.
6.2 Winkel Media
Winkel Media has built a retail-media network inside more than 6,000 Mexican corner stores, running digital signage and using point-of-sale and transaction data to sell targeted advertising to CPG brands[14]. Reported results include sales increases exceeding 40% in the food category on its campaigns, reach of roughly 11 million consumers a month, and plans to add 3,000 to 5,000 stores a year[15]. It works for the opposite reason Juntos+ works: it never touches inventory, transportation or credit at all. It monetizes something that already exists — foot traffic and shelf space — without owning any of the capital-intensive layers that sank the ventures in Section 5. Its constraint is scale, not model: 6,000 stores is a real, profitable base, but a small fraction of the more than 1.1 million storefronts sized in Section 2.
7. The Persisting Gap
Both of these models are proof that important parts of this thesis already work commercially. Neither, on its own, closes the gap described in Sections 3 and 4.
No single, neutral source of supply. Juntos+ solves multi-category ordering — but only within FEMSA’s own portfolio. A retailer using it still has to separately manage every other supplier (snack distributors, cleaning-products wholesalers, produce, and so on) exactly as before, with no way to compare FEMSA’s own price or delivery terms against a competing distributor’s inside the same app, because the platform is built, owned and priced by one of the suppliers competing for that shelf space.
A structural conflict of interest. FEMSA, which operates Juntos+, also owns and continues to expand OXXO[16] — the same professionalized, capital-strong convenience-store competitor identified in Section 3 as squeezing independent retailers’ foot traffic. A retailer-facing platform owned by the parent of its largest organized-retail competitor cannot credibly present itself as a neutral advocate for the independent channel’s collective interest, however well it serves FEMSA’s own commercial objectives.
The communication and insight opportunity remains almost entirely uncaptured. The traditional network reaches close to 100% of the population and represents the majority of consumer-goods sales value (Section 2), yet Winkel Media — the operator proving this layer can be monetized well, with sales lifts brands are willing to pay for — has activated barely half of one percent of the country’s stores. The other 99%-plus of the network’s collective media, promotional and data potential sits idle.
Put together: the gap is not a lack of proof that this channel matters commercially — FEMSA and Winkel Media each prove a different piece of that case. The gap is a platform that is neutral across every distributor and brand, so it can deliver an honest “one place to source everything” without a conflicted owner, and that captures the communication and data opportunity at the scale of the whole network rather than an early foothold within it. Section 8 turns that gap into a set of design principles worth discussing.
8. A Different Way to Think About Serving Retailers
None of what follows is a build plan. It is a set of design principles worth debating, each grounded directly in the evidence from Sections 5 through 7 rather than in a generic playbook — what has actually failed, what has actually worked, and why.
8.1 Stay capital-light: partner, don’t build
Never own stock, trucks or a delivery force. Every unit that moves should move on an existing cash-and-carry wholesaler’s or distributor’s own logistics — the exact asset that Frubana, Merqueo, Udaan and Wasoko/MaxAB each tried to replicate or subsidize, and could not sustain[17], [18].
8.2 Distributors as the platform’s second core audience
Rather than treating distributors and C&C wholesalers as a channel to route around, they should be a primary audience the platform is built to serve — alongside retailers, not beneath them. A distributor’s reach, service quality and portfolio become visible and comparable in ways no individual distributor could achieve alone, giving small and mid-size players a collective way to compete with a vertically integrated giant like FEMSA without needing FEMSA’s balance sheet. Every workflow — order routing, catalog updates, service scorecards — should start from a distributor’s existing processes and judgment, taking seriously the role Udaan’s own retrospective says these players actually play: credit assessor, collection agent, demand forecaster, relationship manager, brand educator, local logistics coordinator and informal risk manager[19], not an algorithm to be substituted for that judgment.
8.3 Transparency by contribution, not by scraping
Drop any notion of scraping distributor prices from public sources. Price and catalog visibility should be a feature distributors opt into and control — publishing their own current price list, promotions and delivery terms because doing so wins retailer volume, not because their listings were taken without consent. That single change turns a likely point of channel conflict into a service the platform can sell to distributors instead of a reason for them to resist it.
8.4 Neutrality and data as the core differentiator
The platform should never take a position — through ownership, pricing or promotion — that favors one distributor or brand over another. This is precisely what neither Juntos+ nor any other single-brand or single-distributor app can structurally offer: a retailer-level, cross-brand, cross-distributor view of what is actually selling, and a market-level, cross-competitor view for brand owners and distributors that a vertically integrated operator can never credibly provide about itself, let alone about its rivals. Genuine neutrality is not a feature among many here — it is the whole differentiator.
8.5 Revenue sequencing: lead with what already works
Launch with the two revenue lines already proven, separately, in this exact market: a retail-media and data layer sold to brand owners, on Winkel Media’s model[20], and a SaaS-style fee charged to distributors for dashboards, service-evaluation analytics and campaign/task administration. Order-aggregation fees and any consumer-facing loyalty or rewards economics should be treated as second-stage revenue, introduced once the network and its data have real scale — the reverse of the sequencing that preceded the capital-heavy failures in Section 5.
8.6 Credit: partner, never underwrite
Any financing offered to retailers should be originated and underwritten by an existing regulated lender or fintech partner, brought in once there is enough transaction and service-evaluation history on the platform to make that partner’s own underwriting easier — never built as in-house lending. This follows directly from Udaan’s own conclusion that app-based credit scoring missed the local, relationship-based nuance a human distributor applies[21], and keeps credit risk off the platform’s own balance sheet entirely, consistent with 8.1.
8.7 Reinventing the field force: from selling to digitizing and retaining
The pre-saler/prompt-saler model should not be eliminated outright — retailers, generally prefer someone else to place the order for them rather than self-serve on an app. What should change is the job itself: a leaner field force of “digitizers” focused on onboarding new stores, training owners on the platform, and — this is the recommendation worth underlining — making retention visits triggered by actual need signals (a store gone quiet on ordering or evaluating, a competitor’s promotion nearby, an unresolved service complaint), rather than blanket fixed-frequency routes regardless of whether a given store needs a visit that week. This keeps the human relationship retailers value while turning field coverage into a data-informed retention tool instead of a blind, fixed cost center — a materially cheaper structure than carrying a full inventory-bearing sales force.
8.8 Data governance from day one
Because neutrality and the data/insights layer are the platform’s core differentiator (8.4), aggregating transaction and consumer data across many distributors, brand owners and retailers has to be built on an explicit data-governance and consent framework aligned with Mexico’s federal data-protection requirements from the very first pilot — not retrofitted once the data layer starts generating revenue. Trust in that neutrality is the asset every other design principle in this section depends on, and it is far easier to build in from the start than to repair after the fact.
9. How This Could Unfold — A Phased Story
Told as a story rather than a project plan, and deliberately paced by proof rather than ambition, expansion here happens along two axes at once: which parts of the solution are live, and which geography they are live in.
Phase 0 — Trust, in one metro area
A small number of anchor cash-and-carry wholesalers and distributors join as founding partners — chosen so early retailers see genuine category breadth rather than one supplier again. Digitizers (8.7) onboard an initial cluster of retailers. The only things live are self-registration, a distributor-input catalog and pricing layer (8.3), and a standardized, distributor-visible service-evaluation loop. No campaigns, no credit yet — the phase exists to build trust and a first honest dataset, not revenue.
Phase 1 — Prove the media layer
Once a retained base of actively ordering and evaluating retailers exists, the retail-media and campaigns module opens to a first handful of brand owners, on Winkel Media’s proven model. This is the platform’s first real revenue, and it is deliberately the capital-light one (8.5).
Phase 2 — Deepen the distributor relationship
The distributor-facing SaaS fee (8.5) is formalized: dashboards, task and campaign management, service-evaluation analytics. Distributors move from pilot partners to paying platform customers, and their renewal and satisfaction become the gate the platform must clear before broadening further — treating them, in practice, as the second core audience described in 8.2.
Phase 3 — Widen sourcing breadth
With distributor trust and a real dataset established, cross-distributor order visibility extends so a retailer can genuinely compare and source across every participating supplier in a category — the “single neutral source” promise from Section 7 — still fulfilled entirely through each distributor’s own delivery fleet, never the platform’s own.
Phase 4 — Bring in credit, through a partner
A financing product is introduced through a partner bank or fintech (8.6), using the transaction and evaluation history accumulated since Phase 0 to support that partner’s own underwriting rather than replace it.
Phase 5 — Scale two ways at once
Expansion replicates the full loop metro by metro rather than only adding stores inside one city — and within each new metro, the platform restarts at Phase 0’s narrower scope before layering later phases back in. Growth is paced by proof in both geography and solution scope at once, rather than attempted everywhere simultaneously — the direct answer to the “replacing too much of the ecosystem too quickly” failure mode documented in Section 5[22].
Across every phase, the digitizer force (8.7) and the data-governance framework (8.8) are treated as fixed infrastructure from Phase 0 onward, not later additions bolted on once the platform is already generating revenue.
10. Questions for Discussion
This paper is meant to prompt a conversation, not end one. Some of the open questions worth putting to that conversation:
Which distributors or C&C wholesalers in a candidate pilot metro would realistically be willing founding partners, and what would it take for them to expose their pricing and service data on a platform they do not own?
How large a retained, actively evaluating retailer base is genuinely needed before brand owners will pay for the media layer, benchmarked against Winkel Media’s own multi-year ramp to 6,000 stores?
Is there a case for partnering with an operator like Winkel Media on the media/data layer rather than rebuilding it — and what would that mean for the neutrality principle in 8.4?
What would make a bank or fintech partner comfortable underwriting credit off this platform’s data, and how much transaction history would that realistically require?
Should a vertically integrated brand owner such as FEMSA be invited to participate as one distributor among many once the platform exists — and can genuine neutrality survive including the largest, most conflicted player as a participant?
What data-governance and consent model would retailers and distributors actually trust, given how sensitive sales and pricing data feel to share with any third party — and who should govern that framework as the network grows?
[2]Quirks Market Research — “Traditional Trade in Mexico: The Hidden Giant of Mass Consumption”: over one million traditional-trade stores, nearly 100% population penetration, about two million people employed, roughly 40% of the mass-consumption industry’s total value, average monthly store sales between 24,000 and 46,000 pesos. https://www.quirks.com/articles/traditional-trade-in-mexico-the-hidden-giant-of-mass-consumption
[9]Vanik — “Why Udaan and India’s B2B wholesale marketplaces struggled”: Udaan’s own retrospective concluded that distributors function as “credit assessor, collection agent, demand forecaster, relationship manager, brand educator, local logistics coordinator and informal risk manager,” a role the platform underestimated when it tried to replace rather than support them; combined marketplace, logistics and credit operations became, in the article’s words, a capital “trap” that thin wholesale margins could not sustain. https://www.vanik.com/blogs/udaan-b2b-wholesale-ecommerce-india/
[10]TechCrunch / TechCabal — Wasoko and MaxAB, two of Africa’s largest B2B e-commerce and buy-now-pay-later platforms for informal retail, completed a merger in August 2024 driven by financial distress, amid layoffs and an employee legal dispute in Kenya over the process. https://techcrunch.com/2024/08/27/wasoko-maxab-complete-merger
[18]Vanik — “Why Udaan and India’s B2B wholesale marketplaces struggled”: Udaan’s own retrospective concluded that distributors function as “credit assessor, collection agent, demand forecaster, relationship manager, brand educator, local logistics coordinator and informal risk manager,” a role the platform underestimated when it tried to replace rather than support them; combined marketplace, logistics and credit operations became, in the article’s words, a capital “trap” that thin wholesale margins could not sustain. https://www.vanik.com/blogs/udaan-b2b-wholesale-ecommerce-india/
[19]Vanik — “Why Udaan and India’s B2B wholesale marketplaces struggled”: Udaan’s own retrospective concluded that distributors function as “credit assessor, collection agent, demand forecaster, relationship manager, brand educator, local logistics coordinator and informal risk manager,” a role the platform underestimated when it tried to replace rather than support them; combined marketplace, logistics and credit operations became, in the article’s words, a capital “trap” that thin wholesale margins could not sustain. https://www.vanik.com/blogs/udaan-b2b-wholesale-ecommerce-india/
[21]Vanik — “Why Udaan and India’s B2B wholesale marketplaces struggled”: Udaan’s own retrospective concluded that distributors function as “credit assessor, collection agent, demand forecaster, relationship manager, brand educator, local logistics coordinator and informal risk manager,” a role the platform underestimated when it tried to replace rather than support them; combined marketplace, logistics and credit operations became, in the article’s words, a capital “trap” that thin wholesale margins could not sustain. https://www.vanik.com/blogs/udaan-b2b-wholesale-ecommerce-india/
[22]Vanik — “Why Udaan and India’s B2B wholesale marketplaces struggled”: Udaan’s own retrospective concluded that distributors function as “credit assessor, collection agent, demand forecaster, relationship manager, brand educator, local logistics coordinator and informal risk manager,” a role the platform underestimated when it tried to replace rather than support them; combined marketplace, logistics and credit operations became, in the article’s words, a capital “trap” that thin wholesale margins could not sustain. https://www.vanik.com/blogs/udaan-b2b-wholesale-ecommerce-india/
Leave a Reply