In CRM as in ERP, every serious vendor will answer yes to almost every line in your requirement matrix. The questions that separate them are somewhere else entirely.
A CRM or ERP tender usually arrives as a spreadsheet. Three hundred requirement lines, a column for each vendor, and a scoring key running from zero to three. Six weeks later, every vendor has scored between 2.6 and 2.8, the differences are inside the margin of interpretation, and the decision quietly falls back on whichever demo felt best and whichever price looked lowest. The matrix did not decide anything. It made the decision look defensible after the fact.
That is not a criticism of the people who build these documents. A requirement matrix is a reasonable instinct: write down what we need, ask the market whether they can do it. The problem is that in a mature software category, they can. Contact management, pipeline stages, quotes and dashboards in CRM; ledgers, order handling, inventory, procurement, period close and consolidation in ERP; an integration layer and some form of AI assistance in both — the systems a Danish company is likely to shortlist will do all of it, well enough. Asking whether a capability exists no longer separates anything. Asking how it behaves in your business still does.
Settle the decision before you ask the market
Most of the value in a CRM selection is created before a single vendor is contacted, and it consists of four things written down in plain language.
The first is the sales process as it will actually be run — not as it is drawn on the strategy slide. Stages with names the sellers recognise, and exit criteria for each: what has to be true before an opportunity moves forward. If the organisation cannot write this down, no system will supply it. A CRM configured around an undefined process becomes an expensive activity log within two quarters. In ERP the equivalent question is larger and harder: how far the organisation intends to standardise, and where local variation is genuinely justified. That is a decision about the business, not about software, and no vendor can take it for you.
The second is the data model. What is a customer: a legal entity, a group, a site, a person? Where does the account master live — in the ERP or in the CRM — and which way does the record flow when the two disagree? For ERP, ask the same question of the item, supplier and cost structure, and of the hierarchy the group consolidates through. This single question determines more about the integration cost, and more about whether management will trust the pipeline numbers, than any feature comparison.
The third is the definition of the numbers the management team intends to run on. Weighted pipeline, forecast categories, win rate measured from which stage, revenue recognised when — and in ERP, the chart of accounts, the reporting hierarchy, and how long the close is allowed to take. Vendors will show you beautiful dashboards built on their definitions. Bring your own.
The fourth is ownership after go-live: who administers the system, with how many hours per week, and who decides what gets changed. Neither a CRM nor an ERP is a project that ends. If nobody is named, the system drifts, and the drift is usually visible within a year — faster in CRM, more expensively in ERP.
Questions a vendor cannot answer with yes
Once the frame is set, the tender itself can be built around evidence rather than assertion. The most useful shift is from written answers to demonstrated behaviour: instead of asking whether the system supports your quoting process, hand each vendor the same script — your account structure, your product hierarchy, three real opportunities with their complications intact, or for ERP an order carried through to invoice and a period close on your own data — and ask them to run it live, in that order, in the same time box. What separates vendors is not whether they finish. It is what they had to configure, work around, or defer to a partner to get there, and whether their consultant could do it in front of you.
Cost deserves the same treatment. A licence price per user per month is the smallest part of the picture. Ask for a five-year total that names implementation effort by role, integration build and maintenance, sandbox and environment costs, the administration capacity the vendor assumes you will hold internally, and the price of the tier you will realistically be on in year three rather than the one you enter at. Ask explicitly what triggers a step up in cost — contact volume, transaction volume, report complexity, permissions depth, additional legal entities — because that is where both CRM and ERP business cases most often come apart.
Then ask two questions that professionals answer well and salespeople answer badly. First: what would you take out of this scope if the timeline were half as long? Second: describe a client of comparable structure where this did not go smoothly, and what you changed as a result. The quality of those two answers tells you more about the delivery team you will actually get than any reference visit arranged by the vendor. On references, ask for a list and choose from it yourself.