The Case of the Missing Value: What No One Audits Between Giants and Startups

The crime scene: a lost customer at 09:17

09:17 a.m., Monday. A customer closes their account forever at a traditional bank using the app of a neobank they opened the night before in eight minutes. The bank will report “normal portfolio churn.” The neobank will celebrate “user growth.”

Nowhere in the reports does what I’m looking for appear: the vanished value.

Because neither the bank nor the startup is measuring the real cost of that move: fees that are lost, risk that becomes concentrated in a still‑fragile player, data that exits a regulated system into a semi‑fragmented one, service expectations that skyrocket without a solid economic model behind them.

That’s the kind of discrepancy that triggers my auditor’s alarm. I don’t investigate accounting fraud; I chase narrative fraud. And the dominant narrative says incumbents and startups are fighting to “win the future.” When I follow the trail of money and data, I find something else: accounting holes in value that nobody is auditing.

We’ll treat four industries—financial services, retail, health and mobility—as a set of crime scenes. I’ll always use the same instrument: a forensic matrix of business model, technology/data and user experience. Not to repeat the story of “the slow versus the agile,” but to answer a more uncomfortable question:

When a sector is digitalized, how much value is truly created, how much is merely shifted from one player to another, and how much simply disappears in the process?

The X‑file: a forensic matrix to follow the value trail

Innovation reports usually talk about competitive advantages. I prefer to talk about value ledger entries: who generates it, who captures it, who finances it and who bears it in the form of risk.

To structure the investigation, I use a three‑axis matrix that can be applied to any sector.

1. Business model: the visible ledger

Here I ask:

• Who pays, for what, how and when?
• Is revenue recurring or purely transactional?
• Is the risk on your own balance sheet or outsourced (to partners, users, regulator)?
• Does diversification protect or dilute focus?

2. Technology / data / AI: the hidden accounting

This axis answers:

• What share of operating costs is legacy that no one dares to write off?
• To what extent does the tech architecture allow integration, experimentation and scaling without rebuilding everything each time?
• Do the data really serve to make decisions, or only to decorate slide decks?
• Who truly controls the sector’s “data ledger”?

3. User experience: the emotional cash flow

No balance sheet can withstand a toxic UX. Here I analyze:

• Where is the real friction concentrated (not the friction marketing claims to have solved)?
• Is omnichannel truly coherent or just a collage of unreconciled channels?
• Does personalization create value or just squeeze margins with discounts?
• Does design reduce anxiety and time or merely add gloss to old processes?

With this framework, let’s move on to the four crime scenes.

Scene 1: Financial services / Fintech – Money moves, risk stays put

The invisible conflict in banking

Traditional banks present themselves as prudent fortresses. Fintechs as customer liberators. The money trail says something else: regulatory, reputational and systemic risk remains on the incumbent’s side, while part of the margin and emotional relationship escapes to startups.

Business model: old fees, new promises

• Incumbents (traditional banks, insurers)
  Diversified income: loan interest, product fees (cards, funds, insurance), advisory services. Broad B2C and B2B portfolios, often with full business units dedicated to corporate and investment banking.

  Diversification stabilizes results but makes focus and rapid iteration harder. Shutting down an obsolete product line means touching structures, bonuses and internal politics.

• Fintech startups (neobanks like Nubank, Revolut; players like Chime, BaaS, aggregators)
  Very focused B2C or B2B2C models: digital accounts, payments, remittances, consumer credit, or “banking‑as‑a‑service” where a backend bank provides the balance sheet and regulatory cover. They generate revenue via interchange, premium subscriptions, low but scalable fees and, increasingly, monetization of behavioral data.

Iteration speed is asymmetric: a fintech ships a new feature in weeks; a bank, in quarters. But almost nobody asks: who absorbs the cost of mistakes when something goes wrong? The answer is clear in the regulatory record: usually the incumbent.

Technology and data: legacy as hidden liability

• Architecture
  Banks: core banking systems on mainframes, on‑premise, point‑to‑point integrations, middleware layers stacked like technological archaeology. Each new mobile app is really another layer of makeup on a costly structure.
  Fintechs: cloud‑native architectures, microservices, API‑first, heavy use of SaaS.

• Analytics and AI
  Fintechs use AI/ML for alternative risk scoring, fraud detection and offer personalization. Banks do too, but tied to historical data silos that are hard to reconcile. The potential of those data is enormous, but the cost of putting them in shape is astronomical.

• Integration and governance
  With open banking, many startups access bank data via APIs. The bank funds security, compliance and traceability; the fintech monetizes the interface. Data governance is usually better documented at incumbents, simply because the regulator forces them to.

User experience: friction as hidden tax

• Incumbents
  High levels of friction in customer onboarding, mortgage applications or complaints. Onboarding that requires branch visits, physical documents or poorly designed KYC processes. Fragmented omnichannel: what a client starts in the app doesn’t always continue smoothly at the branch or call center.

• Fintechs
  Fully digital onboarding in minutes, polished mobile‑first interfaces, real‑time notifications, almost everything self‑service. Friction drops so much that customers sometimes take risks they don’t understand (crypto, BNPL, leverage) just because the button is big and green.

Forensic snapshot: who seems to win and who pays the bill

Forensic conclusion for the sector: the story says fintechs “beat the bank.” The ledger says much of the risk and structural cost remains on the incumbent’s balance sheet, while emotional and data value migrate to startups. The value that disappears is in the middle: over‑promises of free and convenient services that nobody yet knows how to sustain long term.

Scene 2: Retail and e‑commerce – Margin leaks out through the last mile

The invisible conflict in retail

Retailers and supermarkets complain that “platforms take the customer.” Platforms complain that “logistics eats the margin.” The receipt and the logistics cost sheet reveal the crime: constant erosion of unit margin disguised as customer convenience.

Business model: from selling products to subsidizing delivery

• Incumbents (brick‑and‑mortar retail, big chains, supermarkets)
  B2C model centered on volume and turnover. Revenue from direct sales with tight margins, sometimes complemented with in‑store advertising, private labels and related financial services (store cards). High diversification across product categories and store formats.

• E‑commerce and marketplace startups (Glovo‑type convenience marketplaces, Wallapop‑type C2C, vertical D2C)
  Marketplace, B2C, C2C or B2B2C models. Revenue from intermediation fees, merchant subscriptions, promotional services (ads) and, in some cases, customer subscriptions (free shipping, exclusive deals). Business focused strongly on niches: urban last mile, second‑hand, specific categories.

Iteration in pricing, assortment and services is much faster at startups. But their P&Ls hide a key fact: the last mile is subsidized. Each fast delivery with an apparently low cost is an accounting entry that, in many cases, still doesn’t add up.

Technology and data: inventory as blind spot

• Architecture
  Traditional retail typically runs on on‑premise ERPs and inventory systems, legacy POS and online channels bolted on afterwards. Startups use cloud platforms, modern OMS, real‑time routing and SaaS solutions for payments, fraud and customer service.

• Analytics and AI
  Startups use data for dynamic pricing, recommendation engines, route optimization, fraud detection and campaign personalization. Many incumbents are only just beginning to coherently integrate physical store and online data.

User experience: customers buy experiences, not just products

• Incumbents
  In‑store experience depends on staff, queues and layout. Online is often a separate channel, with unsynchronized stock and heavier returns processes. Friction: opening hours, travel, lack of personalization.

• Startups and platforms
  UX focused on convenience: shopping from the couch, order tracking, reviews, price comparison. Omnichannel is better resolved on the digital side, even if they depend on third‑party warehouses or stores. Interfaces heavily optimized to reduce time to purchase.

The margin erosion picture

Forensic conclusion for the sector: customers gain convenience, but the whole system funds that convenience through compressed margins. The value that disappears is obvious: sustainable profitability across the entire chain. Neither retailers nor many startups have yet closed the “last mile” entry in a healthy way.

Scene 3: Health / Healthtech – Sensitive data, opaque incentives

The invisible conflict in health

In healthcare, the dominant narrative speaks of “empowering the patient.” My auditor’s experience says the patient is still the last to know who accesses their data, who bills for them and who trains algorithms with them.

Business model: from medical acts to continuous data flows

• Incumbents (hospitals, clinics, insurers)
  Revenue from medical acts, stays, diagnostics and insurance policies. B2C and B2B models (companies buying coverage). High diversification in specialties and auxiliary services. Incentives often linked to volume of activity, not necessarily to health outcomes.

• Healthtech startups (platforms like Zocdoc/Doctoralia, telemedicine, connected devices)
  B2C, B2B2C and B2B models: appointment intermediation, subscriptions for health professionals, practice management SaaS, telemedicine, remote monitoring solutions. Revenue from subscriptions, transaction fees, software licenses.

Product iteration speed is very different: a startup launches a new teleconsultation feature in weeks; a hospital needs months to change a protocol.

Technology and data: interoperability as repeat offense

• Architecture
  Incumbents depend on legacy electronic health records, fragmented by department, with little interoperability between hospitals and insurers. Often on‑premise, with minimal integrations.
  Startups operate with cloud, APIs, mobile apps and wearables, but hit integration walls when they try to plug into the clinical core.

• AI and analytics
  Startups apply AI to automated triage, reminders, image reading, chronic patient monitoring. They have fewer historical data, but cleaner and more actionable.
  Hospitals hold decades of clinical data, but suffer from quality issues, heterogeneous formats and strong legal constraints.

User experience: from crowded corridors to video calls

• Incumbents
  Experience defined by waiting rooms, opaque prices, redundant admin processes, limited communication. High emotional friction: anxiety, sense of lack of control.

• Healthtech
  Digital onboarding, appointment booking in seconds, reminders, access to basic records, chat with professionals, remote follow‑up. The experience feels more human, even though it’s mediated by a screen.

Health: the balance of value and risk

The narrative sells convenience and access. The data trail shows something else: a silent transfer of sensitive data to platforms with still‑unclear business models. Who will be able to train AI models with those data, under what controls and with what value‑sharing schemes is something that appears in nobody’s “income statement.”

Forensic conclusion for the sector: incumbents retain the regulatory and clinical burden; startups capture user sympathy and data interfaces. The missing value on the balance sheet is transparency about who monetizes each data point and under what responsibility.

Scene 4: Mobility and logistics – Promised time, saturated capacity

The invisible conflict in mobility

Mobility and delivery apps promise almost infinite speed and availability. Route sheets, labor contracts and logistics income statements tell a different story: a system pushed to its limits, where the real cost of time is externalized onto drivers, couriers and traditional operators.

Business model: from selling transport to selling minutes

• Incumbents (transport fleets, traditional logistics firms)
  B2B models: transport contracts, warehousing, supply chain management. Relatively predictable revenue, tight margins, medium‑/long‑term contracts.

• Mobility and logistics startups (Cabify, last‑mile platforms like Glovo, transport marketplaces)
  B2C and B2B2C platforms: mediating between supply (drivers, couriers) and demand (end users, restaurants, e‑commerce). Revenue from commissions on transactions, dynamic pricing, premium services.

Iteration in fares, promos and services is very fast at startups. But that dynamism often rests on a traditional operational base: warehouses, roads, logistics operators, transport regulations.

Technology and data: surface optimization, deep rigidity

• Architecture
  Incumbents: legacy TMS (Transportation Management Systems) and WMS (Warehouse Management Systems), limited real‑time visibility, weak integration with end customers.
  Startups: cloud platforms, dynamic assignment algorithms, AI‑optimized routes, mobile apps for drivers and customers.

• Data and AI
  Digital platforms collect geolocation, timing and demand pattern data. This lets them optimize prices and wait times. However, using these data to coordinate with physical infrastructure (ports, hubs, warehouses) remains limited.

User experience: the “progress bar” effect

The user sees a progress bar showing when their car or order will arrive. That bar is gold: it reduces anxiety, builds trust and loyalty.

• Incumbents
  Offer partial tracking, often only for B2B shipments. Customer service based on calls, little transparency.

• Startups
  Offer real‑time visibility, on‑the‑fly address changes, continuous ETAs. Interfaces designed to minimize the feeling of waiting.

Forensic conclusion for the sector: users gain perceived control; the system gains hidden complexity. The unseen value is the capacity to absorb shocks (fuel price hikes, labor regulation, congestion) without breaking the economic model of platforms and operators.

The recurring pattern: who has structural advantages… and responsibilities

Across sectors, the money and data trail shows the same pattern.

Structural advantages of incumbents (and their flip side)

• Scale, brand and regulatory trust.
  This lets them absorb shocks but anchors them to slow processes.
• Access to cheaper, steady capital.
  Enables big transformation bets, but makes it hard to “kill” sunk projects due to pride or politics.
• Massive historical data.
  Huge AI potential, saddled with quality and governance liabilities.
• Physical distribution networks (branches, stores, hospitals, hubs).
  Hard‑to‑replicate competitive advantage, expensive to maintain.

Structural advantages of startups (and their accounting limits)

• Modern tech architecture, no legacy.
  Scaling is cheaper, but any design error is amplified.
• Digital talent and experimental culture.
  High agility at the cost of churn and weak organizational memory.
• Initially lower fixed costs.
  But much of the structural cost lies with third parties (cloud infra, banks, hospitals, logistics operators) who can change terms.
• Extreme focus on niches.
  Great for early traction, risky when the niche saturates or commoditizes.

Convergence: when the numbers stop adding up on both sides

Incumbents trying to act like startups

We see banks launching internal neobanks, retailers opening digital “labs,” hospitals creating innovation units, logistics operators buying marketplaces. In the accounts, this appears as rising digital CAPEX and OPEX. In the narrative, as “transformation.”

Forensic reality:

• Many labs are cost centers with no clear P&L.
• Core tech transformation moves at a pace incompatible with the expectations sold to the market.
• Startup acquisitions sometimes buy tech; often they buy mainly talent and narrative.

Startups discovering they can become incumbents too

Several fintechs are now applying for full banking licenses. Logistics marketplaces end up buying their own fleets. Health platforms open physical clinics or sign exclusivity deals with hospital chains.

Each of these moves adds structural liabilities to the balance sheet and reduces the lightness they originally sold as an advantage.

In the system‑wide ledger, convergence looks like this:

• Incumbents multiply their digital complexity.
• Startups accumulate physical, regulatory and labor obligations.

The question almost nobody asks is: is the end customer willing to pay the full price of that convergence, or are we expecting someone else to absorb the difference?

The reckoning: strategic shifts that can’t be postponed

If we accept that there is “vanished” value—not destroyed, but mis‑recorded—the task isn’t just to innovate, but to reconcile the ledgers: economic, technological and experiential.

For incumbents: three auditor’s shifts, not marketing ones

1. Clearly separate experiment P&Ls from core P&Ls
   Stop hiding innovation labs inside traditional units. Every “startup‑like” initiative needs its own income statement, targets and time horizon. Without this, the organization never learns what works or at what cost.

2. Treat tech legacy as an explicit liability
   Quantify the real cost of maintaining old systems and define an amortization and replacement plan. As long as legacy is a “fuzzy cost,” it will always lose out to flashy projects.

3. Measure user experience as a trust cash flow
   Bring real friction metrics (wait times, process steps, error rates by channel) into executive dashboards. UX is not aesthetic; it’s a driver of churn, reputational risk and regulatory pressure.

For startups: from narrative to grown‑up accounting

1. Model the full cost of convenience
   Build into the economics everything that’s currently subsidized (last mile, cashback, near‑zero commissions, 24/7 support) and define profitability thresholds by cohort, not just total users.

2. Build explicit strategic relationships with incumbents
   If your model depends on banks, retailers, hospitals or logistics operators, turn them into strategic partners with clear agreements on data, risk and value sharing. Hidden dependence is a systemic risk.

3. Design UX with the regulator in mind
   Ultra‑frictionless experiences in regulated sectors eventually attract supervisory scrutiny. Visible safeguards (risk explanations, usage limits, checks) are an investment in your social license to operate.

The big hidden ledger: who audits systemic value

This tour through four crime scenes makes one thing clear: the “incumbents vs startups” debate is misframed. The real clash is between the innovation narrative and the system’s ledger.

In that ledger, the relevant questions aren’t:

• Who has the prettiest app?
• Who raises the biggest round?
• Who closes more branches?

But:

• Who bears the cost of tech failures when they hit millions?
• Who guarantees service continuity when the business model isn’t yet profitable?
• Who controls and audits data use when it changes hands three times before generating a euro?

Today, many of those answers default to: whoever has the bigger balance sheet, more regulation and more to lose. In other words, the incumbents. Meanwhile, startups capture attention, data and, sometimes, part of the margin.

Real digital transformation won’t be the one that turns everyone into platforms, but the one that subjects value creation and destruction in each sector to serious audit. Until we have a “systemic income statement”—one that accounts for transferred risk, exposed data and resilience lost or gained—we’ll keep celebrating new apps without knowing whether, as a system, we’re getting poorer or stronger.

My hypothesis as an auditor is simple:
once economic, technological and customer‑experience ledgers start to match, we’ll stop talking about “giants” and “startups.” We’ll talk about those who record all the value they create and those who, unknowingly, still leave it off their books.

References

1. Advanced Factories, “Las startups industriales apuestan por la IA, conectividad 5G, robótica e IoT para mejorar la productividad de las fábricas” (2023).
2. Steve Blank, definition of a startup as a temporary organization designed to search for a repeatable and scalable business model. Quoted in: Spanish Wikipedia, “Empresa emergente.”
3. Cámara Valencia – TIC Negocios, “Startup vs empresa tradicional: diferencias clave y conceptos básicos.”
4. Hipatia, “¿Es lo mismo emprender que fundar una startup?”.
5. Sector context and examples of incumbents vs startups according to comparative analysis provided (financial services, retail and e‑commerce, health, mobility and logistics).