Someone walked out of that meeting with a one-pager. Two engineers, six months, full control over the stack. The receipt system, the QR flows, the data capture layer, all built in-house, exactly the way the business needs it. The number on the page looked smaller than any vendor quote they had seen.
Eighteen months later, they had a receipt sender. No progressive profiling. A 9% email capture rate. And a compliance gap that opened when Germany tightened KassenSichV fields in January 2024 that nobody had budgeted to close.
The question was never whether they could build it. It was what they were actually deciding to build.
What You’re Actually Deciding to Build
The estimate almost always covers layer one. Layers two through six are what turn six months into eighteen.
The POS integration layer is what the IT team pictures: a connector between the register and the receipt system. Most mid-market retailers run two or three different POS systems, and every POS update requires regression testing. This layer is not built once. It is maintained continuously, on your team’s time.
The European fiscal compliance engine is what most in-house estimates miss. German KassenSichV requires TSE serial numbers, check values, ERS serial numbers, and consecutive signature counters on every receipt, with fields that tightened in January 2024. France’s AGEC law specifies digital receipt requirements in force since August 2023. Italy mandates digital receipts for large retailers from January 2027. These are not static. Each update becomes an engineering sprint on your calendar, not a vendor’s shared roadmap. In-house teams typically price this function at zero.
The GDPR consent and opt-in architecture is not a checkbox. It is an auditable record, per customer, per market, demonstrating to a regulator exactly how consent was obtained for any given profile. Built from scratch, maintained as regulation evolves.
The progressive profiling system connects a first QR scan to a third store visit to a loyalty enrollment, building anonymous device profiles into identified, opted-in customer relationships without requiring anything upfront. Most in-house roadmaps treat this as phase two. Phase two rarely ships on schedule.
The post-purchase campaign engine converts the receipt from a document into a channel: segmentation, automations, wallet pass issuance, personalised content. Not a static PDF.
The CRM/CDP integration layer keeps data flowing into the systems the retailer already has and rebuilds those connections when the stack changes. It will change.
The IT team estimated the cost of a receipt sender. The actual decision was whether to build and maintain all six layers of a full in-store customer engagement technology stack, from scratch, in-house.
The Costs That Don’t Appear on the Estimate
The build advocate in the room is comparing vendor licensing to two senior engineers for six months. That comparison excludes most of the real cost.
The maintenance bill starts on launch day. Forrester’s Total Economic Impact analysis (2024) puts annual software maintenance at 15 to 20% of initial build cost, with 78% of lifetime software TCO accruing after launch. For a system that cost €200,000 to €350,000 to build, that is a permanent engineering allocation of €30,000 to €70,000 per year, before any new development. The system that shipped in month eighteen is not finished. It is starting.
The opportunity cost is the number most estimates omit. Take a mid-market retailer with 20 stores and 500 daily transactions per store: 10,000 transactions per day. A well-deployed platform achieves a 50% customer identification rate, producing 5,000 identifiable customers per day. The delta between an in-house go-live at eighteen months and a vendor deployment at six weeks is approximately 470 days. At 5,000 identifiable customers per day, that is 2.35 million missed identification opportunities: not a revenue projection, but the customer database that will never be built. Run that calculation for your own numbers.
The compliance monitoring function is priced at zero in most in-house estimates. Reading German implementing decrees, tracking Italian legislative timelines, converting regulatory change into engineering sprints before a deadline becomes an incident: this is a standing function, not a one-time task. For a retailer operating across three European markets, it covers three regulatory environments that do not move in sync.
The honest TCO comparison is vendor licensing against the full build cost, plus annual maintenance at 15 to 20% of that cost, plus opportunity cost, plus the compliance monitoring function. BCG’s 2024 analysis of over 1,000 companies across 59 countries found that more than two-thirds of large-scale technology programmes are not delivered on time, within budget, or to defined scope. That figure reflects what happens when the estimate covered one layer and the system required six.
When Building Is Genuinely the Right Call
Two situations exist where building is the right answer. Both deserve honest consideration.
The engagement stack is your actual competitive differentiator. Some retailers have built a proprietary in-store experience where the physical journey is the product itself: a specific, proprietary engagement architecture that no vendor could replicate and that creates structural advantage over competitors. If the data capture layer is genuinely part of that differentiation, and the team has the expertise to build and sustain it, building is legitimate. The question to ask precisely: is it the data infrastructure that creates competitive advantage, or the experience built on top of it?
You have a large, dedicated product team with deep POS and compliance expertise already in place. Some enterprise retailers, typically those with 500 or more stores and in-house engineering teams of 50 or more, can sustain a build at the required level. The critical qualifier is already. Building the team in order to build the product is a different calculation, and it belongs in the estimate.
For most mid-market European retailers with 20 to 300 stores, neither condition applies. The engagement stack is not their differentiator. Their product, their brand, their buying experience is. The data infrastructure underneath it is a commodity layer. Commodities are bought, not built.
The Four Questions That Settle the Decision
These questions are not rhetorical. Each one is answerable using information already in the room.
Is this stack your competitive differentiator or the infrastructure underneath it?
If the honest answer is “infrastructure,” that is a buy signal. You do not build your own electricity grid because you need reliable power. If the data capture layer is where competitive advantage genuinely lives, and that claim survives internal scrutiny, the answer changes. Most retailers who work through this honestly find it does not.
Do you have the team to build this today, or would you need to hire and train first?
The estimate should include the cost of assembling the team. Ravio’s 2025 Compensation Trends Report puts the fully loaded annual cost of a senior software engineer in Germany at €111,800 to €130,000, including employer social contributions. Two engineers for six months is €100,000 to €130,000 before infrastructure or integration costs. If those engineers do not exist today, add recruitment, onboarding, and ramp time. A vendor deployment at six weeks versus hiring and then building is a different comparison.
What is the full compliance surface you are building for, and who monitors it as it changes?
List every market the business operates in. For each, name the fiscal and data compliance requirements that apply to in-store data capture. Then name the person on the team with the expertise to implement and monitor those requirements as they evolve. If the answer is “we would figure it out,” that risk belongs in the estimate, not in the project retrospective.
What is the daily cost of not being live?
Run the opportunity cost calculation from the section above for your own numbers. Take last month’s daily transaction volume, apply a 50% identification rate, and multiply by the delta between the estimated in-house go-live and a vendor deployment at six weeks. That is the customer database that will not exist if the build runs long. It belongs in the room before the decision is made.
Most retailers who work through these four questions arrive at the same answer. The framework is not constructed to point in one direction. The data usually does it anyway.
If you are moving toward vendor evaluation, the guide to evaluating digital receipt platforms covers how to compare vendors across the dimensions that matter.
What the Right Platform Actually Delivers (That Build Never Reaches)
In-house builds almost always stall at layers one and two. The first twelve months go to making the receipt render correctly and the POS integration stable. The engagement layer, progressive profiling, post-purchase campaigns, loyalty enrollment from the receipt, wallet pass communication, gets deferred to phase two. Phase two rarely arrives on the original schedule, because the compliance work underestimated in month one is still consuming engineering capacity in month fourteen.
The bought platform starts where the in-house build was heading. The compliance and integration layer is already solved. The first weeks of deployment go to receipt design and campaign logic, not to debugging KassenSichV field validation on a legacy terminal.
The practical difference shows in the capture data. In-house implementations that do ship typically see email capture rates of 9 to 15%, because the profiling logic and consent architecture were under-resourced relative to the receipt delivery layer. Strong platform deployments achieve 60 to 70% email capture rates. That is not a feature difference. It is years of iteration on the exact problem that in-house builds are still solving in year one.
The compliance coverage travels with the platform. When Germany updates a KassenSichV field or Italy issues an implementing decree, the engineering sprint is absorbed across the vendor’s client base. An in-house team absorbs it alone.
What that actually means in practice: someone reads the implementing regulation, interprets which receipt fields change and when, scopes the engineering work, tests across every POS environment in the estate, and deploys before the enforcement deadline. For a vendor operating across multiple European clients, that sprint is shared and absorbed in the platform subscription. For an in-house team, it is unplanned and unbudgeted until it isn’t optional.
This is what switching from an in-house build to a platform typically reveals: the retailers who made that move did not give up control. They redirected engineering capacity from maintaining commodity infrastructure to building the things that actually differentiate them.
This is what refive’s platform delivers in practice: not a receipt sender, but a deployed, compliant customer engagement infrastructure that starts capturing data in week one. Book a demo.
The First Step, Whatever You Decide
Whether the conclusion is build or buy, the first step is the same. Run the opportunity cost calculation for the business. Take last month’s daily transaction volume. Apply a 50% identification rate. Multiply by 365. That number is the size of the customer database the business is either building toward or failing to build.
For most mid-market retailers, the number is large enough that the question is not whether to invest in this capability, but how quickly to be live. The in-store customer engagement touchpoints framework maps what that capability looks like across the full customer journey.
If the number is large enough to matter, the next conversation is about speed to live.
Most should not. The answer depends on three conditions: whether the engagement stack is a genuine competitive differentiator, whether the required team already exists, and whether the full compliance surface has been costed rather than just the receipt sender. For most mid-market European retailers with 20 to 300 stores, the data capture infrastructure is not the differentiator. At 10,000 daily transactions, the opportunity cost of an eighteen-month build represents over 2.35 million missed customer identification opportunities.
A receipt sender alone takes three to four months. The full stack, covering POS integration, European fiscal compliance, GDPR consent architecture, progressive profiling, post-purchase campaigns, and CRM integration, takes twelve to eighteen months at minimum. Most in-house estimates cover the first item, and the gap between that estimate and the full build is where most projects stall or ship incomplete. A bought platform deploys in four to six weeks.
Initial build cost for a mid-market retailer runs €150,000 to €400,000 depending on scope and team composition. Ongoing maintenance then runs 15 to 20% of that build cost annually, according to Forrester’s Total Economic Impact analysis (2024), with 78% of lifetime software TCO accruing after launch. The vendor licensing fee looks different against the full number than it does against the initial estimate alone.
Two genuine cases exist. First, the engagement stack is a core competitive differentiator and the team to build and sustain it already exists. Second, the retailer has a large engineering team with deep POS and retail compliance expertise already in place. For most mid-market retailers with fewer than 300 stores, neither condition applies.
Build TCO equals initial development plus annual maintenance at 15 to 20% of build cost plus compliance monitoring plus opportunity cost of delayed deployment. Buy TCO equals vendor licensing plus implementation plus integration. Most build estimates exclude maintenance and opportunity cost. When those are included, the comparison usually reverses.
Yes, technically. But the compliance surface is larger and more dynamic than most teams price. German KassenSichV required new TSE data fields that tightened in January 2024. France’s AGEC specifications have been in force since August 2023. Italy’s requirements for large retailers take effect from January 2027. Vendors absorb each update across their entire client base. An in-house team absorbs it alone.
Ready to see what a compliant, deployed platform looks like before your team makes this decision? The refive demo takes 20 minutes.
