Every invoice your finance team processes carries information that somebody in the business would pay for. Which customers are slowing down. Which suppliers quietly raised prices while nobody was looking. Which products generate revenue and consume all of it in service cost. The back office already holds this, in transactional detail, going back years. It is treated as a by-product of processing rather than as an asset, and the phrase insights as a service is the industry's attempt to name what happens when that changes. The attempt is mostly overpromised. The honest version of this idea is narrower and considerably more useful than the vendor version, and it has three tiers rather than one.
Three tiers, and only one of them is revenue
Tier one: operational reporting. What was processed, how fast, how many exceptions, how many items are outstanding. Necessary, unglamorous, and a cost. Nearly every back office already produces this and mistakes it for analysis. Tier two: decision support. Information shaped to change a specific commercial decision. Which customers should move to prepayment. Which supplier contract to renegotiate first, with the evidence attached. Which sales channel consumes margin in credit notes and rework. This is where the value actually is, and it is not revenue; it is margin and cash improvement, which is better because you do not have to sell it to anyone. Tier three: external data products. Something a third party pays for, whether a benchmark, an aggregated market view, a supplier scorecard or a data feed a customer wants. Real, occasionally very valuable, and available to far fewer organisations than the marketing suggests. The mistake organisations make is attempting tier three from a tier one foundation, which produces a well-designed product built on numbers nobody trusts.
| Tier | Purpose | What must be established |
|---|---|---|
| Operational reporting | Describe processing and outstanding work | Reliable definitions and activity records |
| Decision support | Change a named commercial decision | Owner, date, trustworthy evidence and benefit measurement |
| External data product | Sell a defined deliverable to a buyer | Buyer demand, data rights, confidentiality and support costs |
Qualitative summary of this article's source text, not a measured outcome or performance estimate.
What tier two actually requires
Three prerequisites, none of which are analytics tools. One definition per thing. A customer, a product, a business unit, a cost centre. If two systems disagree about what a customer is, every insight derived from both is arguable, and arguable insight changes no decisions. Transaction history that is complete enough to be trusted. Not perfect. Complete enough that the finance team will defend the number in front of the sales director who does not like it. A named decision with an owner and a date. This is the discipline people skip. An insight that is not attached to a decision someone must make by a certain date is a report, and reports accumulate until they are ignored. Start from the decision and work backwards to the data; never the reverse. I would go further: the single most reliable way to fail at this is to build a dashboard first. Pick three decisions the business makes badly because the information is hard to assemble, solve those three, and demonstrate what changed in money. Everything else follows from having done that once.
Tier three, and the rights question that stops most of it
External data products are genuinely achievable in specific configurations: a distributor that knows regional demand patterns, a logistics operator that knows lane performance, a payments processor that knows sector spending, a services firm that can benchmark cost structures across dozens of similar clients. Before any of that can be sold, the rights question has to be answered, and it is usually answered badly or not at all. Whose data is it. What do your customer contracts say about use of transaction information. What confidentiality obligations attach to supplier pricing, which is almost always more sensitive than people assume. Is there personal data in it, and would aggregation actually make it non-personal or merely feel that way. If you operate in multiple jurisdictions, the answers differ by country and by contract vintage. The test worth applying before you build anything: would you be comfortable telling the customer or supplier whose data underpins the product exactly what you have done. If the answer is no, the legal analysis will not save the relationship when they find out, and they always find out.
Practical Guidance for Transforming Your Back Office Into a Revenue Center
- Start with three decisions, not a platform. Credit terms by customer, supplier renegotiation priority, and channel or product profitability are the usual first three because the data exists and the money is immediate.
- Attach every insight to an owner and a cadence. Who decides, by when, how often it is revisited. An insight without a decision owner is a report with a nicer chart.
- Fix master data definitions before analysis. One definition of customer, product and entity, agreed in writing between finance and commercial. This is the boring work that determines whether anyone believes the output.
- Measure the benefit in commercial terms. Margin points, days of working capital, recovered leakage, avoided cost. Never in dashboards delivered or reports automated.
- Check data rights before packaging anything externally. Customer contracts, supplier confidentiality, personal data, jurisdiction. Do this before the product design, not after the first customer conversation.
- Move analysts toward the commercial teams. Insight produced in a reporting function and emailed to the business dies in transit. The people who understand the transaction data need to sit in the room where the decision is taken.
- Build one external product for one named buyer. Not a data strategy. One customer who has said they would pay, one deliverable, one price. Generalise only after that works.
- Retire the reports you replace. Every organisation doing this accumulates two reporting estates and pays for both. Deleting the old pack is part of the project, not an afterthought.
The Regional Angle
Four regional realities shape what is achievable here, and the first is the most immediately monetisable. Working capital in this market is expensive and relationship-priced. Banks lend against collateral, personal guarantees and familiarity rather than against evidence, and the cost of a facility frequently reflects how legible the borrower is rather than how good the receivables are. A finance function that can produce a clean, structured, auditable ageing of receivables by customer, with concentration, dispute rates and collection history, is holding exactly the information that converts an expensive relationship facility into a cheaper asset-based one. That is not insight as a product; it is insight as a lower cost of funds, and for a mid-sized regional trading group it is usually the single largest return available from back office data. Second, the credit information infrastructure here has matured quickly. Regional credit bureaux now hold meaningful data on corporate and individual obligors, and combining external credit information with your own payment behaviour history produces a materially better credit decision than either alone. Most regional companies still set customer credit limits by seniority of whoever asked, and the gap between that and an evidence-based limit is real money in a market where receivables run long. Third, the tax transition has done the analytical groundwork for you. Value added tax obliged regional businesses to record transactions in a structured, itemised, reconcilable way for the first time, and the move toward electronic invoicing in parts of the region will extend that to a standardised transaction record with the tax authority as an external validator. The compliance burden was real; the by-product is a clean transaction spine that did not exist three years ago. Organisations that treat their tax data as a compliance obligation only are leaving the analysis unclaimed. Fourth, the political constraint, which is specific to the ownership structure of regional business. In family groups and closely held conglomerates, information is frequently held at entity level deliberately. A general manager who has run a division for twenty years may not want group finance able to see customer-level margin, and the owner may be entirely content with that arrangement. Any attempt to build group-level insight from entity data is therefore a governance negotiation before it is a technical project, and the way it succeeds is by making the first product something entity managers want for themselves, such as their own customers' payment behaviour, rather than something that arrives as group oversight.
The objection worth taking seriously
The first objection is capability and it is fair. Most back offices in this region cannot close the month within ten working days, are reconciling intercompany balances manually, and have master data in a condition that makes any aggregate number a matter of opinion. Telling that function to become a revenue centre is asking an organisation that cannot produce reliable accounts to produce insight, and the usual result is a dashboard programme that consumes eighteen months and produces numbers the commercial teams refuse to accept. The second is that data monetisation mostly does not happen. The literature is full of companies discovering their data is an asset and very short on companies with a material revenue line from it. The buyers are fewer than expected, the price is lower, the product needs continuous maintenance, and the internal cost of supporting external customers falls on a team whose actual job is closing the books. The third is relational, and in this region it is decisive. Selling anything derived from customer or supplier transactions risks a relationship that generates real revenue in exchange for a data product that generates very little. In markets built on long-standing personal relationships and exclusive agency arrangements, that trade is usually a bad one regardless of what the contract permits. All three objections point the same way, which is why the recommendation here stops at tier two for nearly everyone. Decision support requires no external buyer, creates no relationship risk, and pays in margin and cash rather than in a new revenue line. It also has the useful property of being self-financing: the three-decision starting point delivers a measurable improvement in a single quarter, which is what funds the master data work that any later ambition would have required anyway. Organisations that genuinely have a tier three opportunity, and a few do, will be much better placed to pursue it after they have proved they can produce a number their own commercial team will defend.
Common Questions
Do we need a data warehouse first?
No. The first three decisions can almost always be answered from existing systems with an analyst and a spreadsheet. Build infrastructure when repeated manual effort justifies it, not before.
Who should own this?
Finance, working directly with the commercial function that owns the decision. A standalone analytics team detached from both will produce reports rather than decisions.
Can we charge internal business units?
You can, and the more useful discipline is simply attributing the improvement. If a credit decision change reduced overdue receivables, say so in the same terms the business uses for its own performance.
What should we expect over the next twelve months?
Expect enterprise software vendors to keep embedding analytics into finance modules, which lowers the tooling barrier and raises the master data barrier. Expect electronic invoicing developments in the region to make transaction data cleaner and more standardised. And expect the first regional finance functions that can evidence their receivables quality to start negotiating noticeably better facility pricing, which will be the most persuasive business case anyone in this field has produced so far.
Transform Your Back Office Into a Revenue Center — we pick the three decisions your transaction data should already be answering, prove the margin and cash impact in a quarter, and build from there.