Crime Scene: What We Really Lost When Giants and Startups Went Digital

The Hook: Bodies on the Floor of the Data Cathedral

The year that finally forced our hand was 2032.

A European bank collapsed overnight after a silent software update mispriced risk for millions of micro‑loans. The balance sheets looked healthy, NPS was high, the app store reviews were glowing. Yet in three Latin American cities, informal economies imploded within weeks. Local merchants who had shifted from cash to digital wallets saw their credit evaporate. Retailers that had closed physical stores in favor of e‑commerce had nothing left to barter. A public hospital, digitized to the bone, suddenly throttled non‑urgent care because reimbursement algorithms flagged whole neighborhoods as “low value.”

On paper, the models across banking, retail, mobility, health, and education were working exactly as designed. The KPIs celebrated by both incumbents and startups lit up green dashboards.

And yet, value had gone missing.

In 2050 we teach this episode as the “Data Cathedral Crime Scene.” When my students reconstruct it, they always ask the same question: How did nobody see it coming? They carefully contrast Banco Sabadell’s capital strength with Revolut’s growth, Amazon’s marketplace logic with the empty high streets, the healthtech sandboxes with the overrun emergency rooms, the edtech apps with the hollowed‑out campuses.

The uncomfortable answer is this: we were arguing about who was winning while ignoring what, exactly, was being lost.

What follows is not another comparison of “agile startups” and “slow giants.” It is a forensic audit of the missing value across five industries, using your own early‑century facts against you.

The Genesis: How the Perfect Comparison Blinded Everyone

Between 2015 and 2025, a consensus narrative took hold.

• Incumbents were solid but sluggish: banks with marble branches and mainframes; retailers with malls; taxi fleets with radio dispatch; hospitals with paper back‑offices; universities with lecture halls.
• Startups were the insurgents: neobanks like Revolut and N26 delivering fully mobile banking; marketplaces like Amazon scaling without owning stock; apps like Glovo and Wallapop turning urban streets into logistics and resale grids; SaaS firms like Typeform monetizing elegant forms; clean‑energy challengers like Holaluz promising a lighter grid; online florists like Colvin optimizing a fragile supply chain.

Analysts sliced the world into clean axes:

1. Business models — subscription, transaction, freemium, B2B, B2C, B2B2C.
2. Technology — legacy vs cloud‑native, monolith vs microservices, on‑prem vs open APIs, low automation vs data‑driven.
3. Experience — branch vs app, queue vs one‑click, generic vs personalized.
4. Organization — bureaucracy vs squads, risk aversion vs experimentation.
5. Patterns — where startups outrun incumbents, where regulation protects the old guard.

The research you used then made the difference look clean:

• Banco Sabadell investing heavily in physical infrastructure, yet also launching a startup hub and accumulating dozens of startup participations.
• Neobanks like Revolut and N26 winning young customers in the mobile, with transparent pricing and fast features.
• Amazon as the archetype marketplace, turning third‑party sellers into its true payload.
• Glovo stitching together more than 25 countries into one on‑demand network; Wallapop turning local second‑hand into a mobile reflex; Typeform building a global SaaS client base in 190+ countries.
• Holaluz riding the renewable wave with hundreds of thousands of customers; Colvin reorganizing the floral chain.

Everyone framed this as a race: who would adopt subscription faster, scale marketplaces larger, automate more processes, push AI deeper into core operations.

You missed that this wasn’t a race.

It was a crime scene, slowly assembling itself.

The Invisible Conflict: Efficiency vs. Redundancy as a Survival Trait

From 2050, the invisible conflict looks obvious: efficiency vs. redundancy.

• Startups optimized for hyper‑efficiency and scale: low marginal cost, asset‑light growth, data‑driven everything.
• Incumbents carried massive redundancies: branches, warehouses, buses with half their seats empty, hospital corridors that could flex in a crisis, campuses that could host people during floods.

You called the first “innovative” and the second “inefficient.”

What almost nobody asked was: what if those inefficiencies were actually social infrastructure in disguise?

When neobanks reduced the need for branches, they also weakened the last physical financial interface many vulnerable customers had. When Amazon’s marketplace outcompeted local retail, it also removed neighborhood‑level stock buffers that had smoothed shocks. When Glovo and similar platforms rewired mobility and delivery, they injected precarious labor into cities but also made food and medicine flows dependent on a few apps and servers. When healthtech digitized diagnostics and scheduling, waiting rooms shrank—but so did informal social support networks. When edtech turned courses into content streams, campuses looked outdated—until you remembered they were also childcare, community, and slow thinking spaces.

The invisible conflict never showed up on the comparison tables:

• Everywhere you saw “cost savings,” we see fragile monocultures.
• Everywhere you recorded “frictionless UX,” we see the removal of buffers and alternative paths.
• Everywhere you celebrated “high NPS,” we see a population that had not yet encountered systemic failure.

The crime was not that startups disrupted incumbents.

The crime was that both sides, obsessed with each other’s metrics, conspired to remove redundancy from the system.

Let’s reopen the files—sector by sector.

Evidence & Insights: Sector Files from the Age of Optimization

1. Banking / Fintech: The Vanishing Safety Net

You wrote glowing case studies of how neobanks like Revolut and N26 “ate the terrain” of traditional entities in the mobile channel. You noted how banks like Banco Sabadell responded with hubs, venture investments in almost a hundred startups, and digital channels, while still bearing the weight of branches and compliance.

The comparative chart you implicitly lived by looked like this:

Scorecard 2024 – Banking vs. Neobanks

From our 2050 vantage point, the missing line in your table is: Systemic resilience.

When everything shifted to mobile, the local branch—as a place, as a ritual, as slow bureaucracy—looked obsolete. Yet in rural regions, those branches had been:

• last‑resort dispute resolution centers;
• semi‑formal credit committees that knew the actual town;
• fail‑safe access when devices, literacy, or networks failed.

The push to “reduce manual processes” and “increase automation”—which traditional banks embraced in order not to lose margins to fintechs—worked. Until a misconfigured algorithm in the early 2030s denied credit to thousands of micro‑entrepreneurs at once. The physical network was already gone. There was nobody left to notice patterns that dashboards framed as “acceptable churn.”

You correctly highlighted regulation and compliance as obstacles for startups. In practice, they were also the last institutional memory of why certain frictions were installed in the first place—fences built after earlier financial fires. In the name of UX and speed, many of those fences were aesthetically redesigned, then gradually removed.

Hidden truth: By 2024, the banking war you thought was about “digital vs. physical” was actually about “thin convenience vs. thick resilience.” Both sides mostly chose convenience.

2. Retail / E‑commerce: The Great Hollowing of Place

Amazon’s marketplace model embodied your ideal of the digital business: low inventory risk, highly scalable, data‑driven. You admired how Spanish startups like Wallapop and Colvin and global players stitched together supply and demand, flattening search friction.

Your own sources celebrated how:

• Amazon monetized a multi‑sided marketplace, charging commissions.
• Wallapop turned local second‑hand deals into seamless mobile interactions for over 15 million users.
• Colvin bypassed intermediaries, serving “hundreds of thousands of customers” with better prices and UX.

On the surface, the comparison made sense:

Retail 2024 – Winners vs. Losers Scorecard

But what never showed up in your comparisons:

• Stores as cooling centers during heatwaves.
• Malls as unspoken public squares for isolated youth and elders.
• Retail workers as informal social sensing systems for neighborhood decline or unrest.

When the pandemic taught you that local space was precious, you briefly noticed this. Then you went back to optimization. You praised Glovo’s ability to expand to more than 25 countries, lining up alliances with McDonald’s and supermarkets, and admired the diversification of services.

You weren’t wrong about their ingenuity. But you rarely asked: What does it mean when one app intermediates half the food and essentials in a city? When the company is sold (as Glovo was to Delivery Hero), and strategic decisions migrate further away, who holds responsibility when a logistical glitch starves a neighborhood?

Hidden truth: Retail’s crime scene is not just the death of the high street. It is the loss of local resilience and social sensing, sacrificed for algorithmic promotions and last‑mile convenience.

3. Mobility / Transport: From Fleet to Platform—and Into the Dark

The story you told yourselves about mobility echoed the others: traditional operators with vehicles and depots; startups with platforms and apps.

Glovo again is emblematic. It wasn’t a classic mobility player, yet it coordinated thousands of riders, effectively rewriting urban logistics. Historically, fleets and public transport systems carried slack—unused capacity in off‑peak hours, drivers who knew their routes by heart, dispatchers who acted as human control centers.

Platform logic changed the game:

• dynamic pricing;
• crowdsourced fleets;
• minimal idle time as a core KPI.

Great for unit economics, catastrophic for shock absorption.

By the late 2020s, when climate events and blackouts became frequent in some regions, systems optimized for minimal slack struggled. Buses and trains that had been cut “for efficiency” were sorely missed. Courier networks operating near full capacity couldn’t easily flex when entire neighborhoods were underwater or communications went down.

The missing metric across your mobility debates was not “on‑time delivery” but adaptive capacity under stress—how much unused, seemingly wasteful capacity a system kept in reserve.

Hidden truth: Startups taught mobility players to see every idle seat, parked vehicle, or unused depot as a failure. You didn’t realize you were criminalizing the very slack that future crises would demand.

4. Health / Healthtech: Data‑Rich, Fragile Care

You were right to celebrate digital health. From telemedicine to AI‑assisted diagnostics, the potential was undeniable. Your own sources described how digital innovation could improve the sustainability of health systems, and how programs like Boehringer Ingelheim’s “Changing Health” tried to align startups, hospitals, and public authorities.[9]

At the same time, you noted the heavy regulatory burden: licensing, data privacy, rigorous validation.[10] You saw sandboxes as clever tools: safe environments for experimentation.[8]

What you rarely said out loud was this: regulation in health is a memory of past harms. Every cumbersome process often traces back to a catastrophe. Startups, pressed by investors, naturally experienced this as drag.

In healthtech, your usual comparative frame looked like this:

You weren’t wrong. Yet you mostly ignored three silent substitutions:

1. From relationship to data point. A family doctor who knows your context was gradually replaced by layered systems: symptom checkers, teleconsultations, algorithmic triage.
2. From public obligation to SaaS contract. As hospitals outsourced modules to startups, the responsibility to maintain, update, and ethically govern those tools shifted into private agreements.
3. From redundancy to centralization. Regional labs and imaging centers were consolidated because digital logistics allowed it. But this also created single points of failure.

In the 2030s, a widely used triage algorithm—trained on biased data and insufficient edge cases—was found to systematically deprioritize chronic care for certain demographics. It wasn’t designed to be cruel; it was designed to be efficient under average loads.

The forensic audit later showed that the bias had been flagged in early validation tests. But procurement teams and startup founders, under pressure, had classified the anomalies as “statistical noise.” The system performed well on standard KPIs: reduced waiting times, higher throughput, lower costs.

Hidden truth: Healthtech’s great missing value wasn’t privacy (important as it is). It was thick clinical judgment embedded in local relationships, slowly thinned out and automated away.

5. Education / Edtech: Content Everywhere, Cohesion Nowhere

By 2020, you had enough evidence to know that customer experience drove revenue. Reports showed that 89% of companies that excelled in CX outgrew their peers, and that modest increases in retention could drive profit jumps of 25–95%.[11][12][13][14][15] You eagerly imported this thinking into education.

Edtech platforms began to treat students as customers:

• optimizing onboarding funnels;
• tracking engagement with analytics;
• boasting NPS and CSAT as if they were course grades.[11][13][14]

Universities, under financial pressure, followed. They launched online programs, micro‑credentials, and hybrid degrees. Many looked to SaaS stars like Typeform as UX role models: intuitive interfaces, low friction, global reach.[4]

This generated impressive numbers:

• more learners reached;
• better “completion rates” on some platforms;
• new markets for professional upskilling.

What your comparisons rarely quantified:

• Network depth — how many meaningful, lasting relationships learners formed.
• Civic function — the role of campuses as spaces where disagreement and collaboration were practiced in person.
• Embodied skills — experiments, labs, maker practices that didn’t translate cleanly to screens.

The edtech startup you celebrated as “scalable” was often parasitic on the physical institutions it claimed to supersede. It could monetize content and interactions precisely because universities had spent decades building reputations, faculties, and social norms.

When some campuses closed during financial crises in the late 2020s, local communities discovered that they had lost not just degrees, but theaters, meeting rooms, childcare possibilities, festival grounds.

Hidden truth: Edtech did not simply “disrupt” education; it relocated value from shared, slow spaces into private platforms, leaving communities with abundant content but weakened cohesion.

The Strategic Shift: How to Stop Erasing the Wrong Kind of Friction

From your perspective in the 2020s, strategic advice revolved around “how to become more like a startup”: adopt agile, migrate to cloud, launch a corporate venture arm, emulate marketplace models, invest in UX.[6][7][15]

Some of this was necessary.

But if we take the crime scene seriously, the real strategic question for a traditional company in any of these sectors was never just how to digitize—it was how to digitize without deleting its survival functions.

Here is the shift that should have happened.

1. Redefine “Inefficiency” as Potential Resilience

Instead of cutting every branch, warehouse, bus line, hospital bed, or classroom that didn’t meet short‑term utilization metrics, leaders could have asked:

• What shadow function does this asset serve in a crisis?
• Which communities rely on it beyond its revenue contribution?

Actionable moves (2024‑2030, in hindsight):

• Banks: Redesign a subset of branches as resilience hubs—multi‑service spaces with cash, digital help, small‑business counseling, and basic connectivity, justified not only by P&L but by systemic stability.
• Retailers: Treat selected stores as local nodes for community logistics and emergency supplies, integrating with but not subordinated to online marketplaces.
• Mobility operators: Maintain redundancy in routes and rolling stock as an explicit resilience asset, modeled and priced into public contracts.
• Hospitals: Preserve local diagnostic capacity and generalist clinics even when central labs seem cheaper, recognizing their role in early anomaly detection.
• Universities: Protect physical campuses as community infrastructure, opening them more widely to local use.

2. Measure What Actually Keeps the System Alive

Your obsession with NPS, CSAT, churn, and short‑term retention was understandable—data was finally plentiful.[11][13][14][15] Yet your own numbers already hinted that good CX yielded higher revenue growth and that a small retention increase produced huge profit swings.[11][12]

What was missing was second‑order metrics.

You could have tracked:

• Resilience density: number of independent access paths for essential services (financial, food, health, learning) per 10,000 inhabitants.
• Redundancy index: proportion of capacity that could be repurposed in less than 72 hours without catastrophic loss.
• Interoperability score: degree to which services could fail elegantly, handing off to alternative providers.

Instead of asking: how fast can a new user onboard into a neobank?

You could have asked: if this app fails abruptly, how easily can users recover access to their assets through another channel?

3. Design UX for Failure, Not Just for Flow

Startups excelled at making flows smooth: one‑click purchases, instant approvals, frictionless sign‑ups.[4][15] You admired them and copied these patterns.

Yet every friction removed is also:

• one less checkpoint for fraud;
• one less pause for reflection (“should I really take this loan?”);
• one less human capable of spotting anomalies.

Strategic shift: design intentional friction into critical flows.

• Banking: require in‑person or video counseling for high‑risk products, not as regulatory theater but as protective ritual.
• Retail: allow for community review or local advisory boards on certain product categories, deliberately slowing harmful categories.
• Mobility: implement failsafe modes where systems degrade gracefully instead of collapsing when algorithms misbehave.
• Health: keep pathways for patient override and clinician discretion, even when algorithms suggest otherwise.[9][10]
• Education: maintain structured, slow cohorts alongside on‑demand content to foster deep learning and mutual accountability.

4. Collaborate Without Offloading Responsibility

You were already experimenting with corporate venture capital, joint ventures, and open innovation. Banco Sabadell’s investment in nearly a hundred startups is a pure expression of that trend; health sandboxes institutionalized it in regulated sectors.[1][8][9]

The missing clause in many of these arrangements was: responsibility under failure.

A better contract between giants and startups would have:

• Defined shared liability for systemic harms (e.g., biased algorithms, catastrophic outages).
• Mandated joint crisis simulations before scaling critical services.
• Required data‑sharing for resilience, not just for growth—standard protocols so competitors could pick up the slack in emergencies.

Instead of asking startups only to be “compliant,” incumbents could have taught them the unglamorous craft of institutional memory: why certain checks exist, what went wrong in 1987 or 2008 or 2020.

5. Recenter Strategy on “Missing Value” Audits

Every board presentation in 2024 should have ended with a simple forensic question:

If this initiative succeeds on every KPI we are tracking—revenue, cost, NPS, engagement—what might we still be quietly destroying?

Make it explicit. Put it in a table next to the business case.

You did this for risk and compliance in a narrow sense. You almost never did it for societal value.

The Big Picture: The Future Belonged to Those Who Kept Some Things Hard

From 2050, your era looks like a period of heroic ingenuity wrapped around a blind spot.

You knew, with data, that:

• superior customer experience directly correlated with revenue growth;[11][12][15]
• modest improvements in retention generated huge profit swings;[11][12]
• startups, by focusing on UX, agile methods, and data, could outrun incumbents in multiple sectors.[4][6][7][15]

You were right to respect that. You were not wrong to copy some of it.

Where you misstepped was assuming that the only rational direction of travel was towards less friction, less redundancy, less locality, and more centralization. You thought you were shedding dead weight. In many cases, you were shedding connective tissue.

From our vantage point, the organizations that survived the shocks of the 2030s and 2040s did something quietly radical:

• They kept branches, stores, depots, clinics, and campuses not as nostalgic ornaments but as multi‑purpose anchors.
• They embraced digital platforms and data analytics—yes—but used them to map and reinforce local resilience, not erase it.
• They adopted startup‑like experimentation while honoring the institutional memory encoded in regulation and “slow” processes.[7][8][9][10]

In other words: they treated every optimization project as a potential crime scene and ran a forensic audit before the harm.

Your manuals spoke of agility, disruption, blitzscaling. The unspoken assumption was that history rewards those who move fastest.

From 2050, the record shows something more nuanced: history rewarded those who knew when not to move fast. Those willing to say: this branch stays, this bus runs half‑empty, this doctor takes the extra five minutes, this teacher keeps a physical classroom—because when the system shakes, these will be our lifelines.

If you’re still reading this from the 2020s, you’re not late.

But you don’t need another playbook on “how to become more like a startup.” You need a forensic habit:

1. Interrogate your own success metrics.
2. Map the redundancies you’re about to cut.
3. Ask who depends on them when the dashboards go dark.

That’s the missing value. Guard it.

References

1. CincoDías (El País). “Banco Sabadell alcanza las 96 startups invertidas y lanza un nuevo ‘hub’ para el ecosistema emprendedor.” 2 Dec 2024.
2. El País Economía. “La gran batalla financiera se libra en el móvil: así le están comiendo terreno los neobancos a las entidades tradicionales.” 14 Dec 2024.
3. Wikipedia (es). “Modelos de comercio electrónico.”
4. Qonto Blog. “15 ejemplos de startups exitosas que inspiran”: Glovo, Wallapop, Typeform.
5. Emprendedores.es. Casos de éxito: Holaluz, Colvin.
6. Infoautonomo.es. “Cómo las startups están redefiniendo los modelos de negocio en el sector financiero.”
7. Realidadeconomica.es. “El impacto de las regulaciones gubernamentales en las startups.”
8. Wikipedia (es). “Sandbox regulatorio.”
9. El País Sociedad. “Salud e innovación digital, un binomio de éxito.” 31 Mar 2025.
10. Fivevalidation.com. “Conociendo los retos regulatorios de los startups de dispositivos médicos y tecnología de la salud.”
11. Asociación DEC. “Métricas clave de la experiencia del cliente que su empresa debería controlar.”
12. PuroMarketing. “Las empresas con buena experiencia de cliente crecen más: estas son las razones.”
13. Lukkap. “Customer engagement vs customer experience.”
14. EADEA. “Experiencia del cliente: métricas y claves.”
15. McKinsey & Company. “The three building blocks of successful customer-experience transformations.”