When someone asks “what does the CRM cost?”, the honest answer is a range with assumptions attached. Subscription fees are the visible slice. A total cost of ownership (TCO) model puts the rest next to it so that comparisons, renewals and budget requests rest on something sturdier than the quote.

You do not need specialist software. A spreadsheet with clear assumptions and an owner will do the job, as long as you keep it alive after the purchase decision.

Choose the Time Horizon and Stick to It

Three years is a common choice because it matches typical contract terms and is long enough for one-off costs to look proportionate to recurring ones. Whatever you pick, use the same horizon for every option you compare. A model that compares one vendor’s three years against another’s one year is worthless.

Structure the Model in Layers

Group costs into categories so that nothing hides inside “other”:

1. Platform fees. Licences by type, add-on modules, sandbox environments, premium support tiers, storage and API-related charges where applicable.
2. Implementation. Configuration, data migration, integration build, testing, training development.
3. Integration running costs. Middleware subscriptions, monitoring, and maintenance effort.
4. People. Administrators, analysts, developers, and the share of time managers and trainers spend on the system.
5. Change and adoption. Training delivery, documentation, internal communications, productivity dip during rollout.
6. Governance and risk. Security reviews, audits, backup arrangements, and contingency.
7. Exit and decommissioning. Often ignored, but real, such as parallel running of the old system.

Pricing and packaging change often, so take figures from current vendor quotes rather than memory, and label each assumption with its source and date.

One-Off vs Recurring, Then Year by Year

Split each line into one-off and recurring, and place it in the year it falls. Implementation costs cluster in year one. Licences grow if headcount grows. Integration maintenance continues indefinitely.

Cost Layer	Type	Year 1	Year 2	Year 3	Source of Estimate
Platform fees	Recurring	Quote	Quote + growth	Quote + growth	Vendor quote
Implementation	One-off	Heavy	Light	Light	Partner estimate
Integration running	Recurring	Moderate	Moderate	Moderate	IT estimate
Internal people	Recurring	Higher	Steady	Steady	HR and finance rates
Training and change	Mostly one-off	Heavy	Refresh	Refresh	Project plan
Exit/contingency	One-off	None	None	Reserve	Risk review

Fill the cells with your own numbers and note the assumption behind each. Keep the table’s structure but let your data drive the content.

Count the People Honestly

Internal time is the line most often left at zero because it does not come from a separate invoice. But an administrator who spends half their week on CRM configuration is a cost whether or not it appears in the project budget. Estimate effort in hours, apply a loaded rate agreed with finance, and show it explicitly. If you refuse to count internal time, a cheaper-looking option that demands more administration will win unfairly.

Model Growth and Scenarios

Build in the changes you expect: more users, more data, more integrations. Then add a low and a high scenario by varying the two or three assumptions that matter most, usually headcount and implementation effort. A model with a single answer pretends to a precision it does not have. A model that says “this lands between X and Y, and here is what moves it” is far more useful to a decision maker.

Be careful with benefits. A TCO model deals with costs. If you also want to argue return, keep benefits in a separate section with their own assumptions, and never treat soft estimates as certain.

A Hypothetical Example

Suppose a company with a modest sales team compares two platforms. Option A has a lower subscription but needs middleware and a part-time developer to maintain a custom integration. Option B costs more per seat but includes a native connector for their billing system.

In the model, Option A’s subscription advantage is partly offset by integration running costs and internal developer time. The totals end up closer than the quotes suggested, and Option A looks riskier in the high scenario. The finance lead does not need the model to name a winner, only to show which assumptions the decision hinges on.

Keep It Alive

After purchase, replace estimates with actuals each quarter. The variance tells you where your estimating is weak, and it gives you a ready-made baseline for the next renewal negotiation. Assign an owner and a review date, or the model will become a historical artefact.

Frequently Asked Questions

How detailed should the model be? Detailed enough that each line has an owner and a source, but not so detailed that nobody updates it. Fewer, well-understood lines beat dozens of guesses.

Should we include sunk costs? No. For a forward-looking decision, include only costs you can still affect. Keep sunk costs in a note for context.

How do we treat discounts that may not renew? Model the list or expected renewal price for later years, or show both the discounted and undiscounted views, so a first-year discount does not mask year-two reality.

Who should build the model? Finance and the CRM owner together. Finance supplies rates and accounting treatment, and the CRM owner supplies technical and operational estimates.

Does TCO replace a business case? No. It addresses cost only. A business case also needs benefits, risks, and strategic fit.

Next Step

Create the seven-layer skeleton in a spreadsheet today, fill what you know from current contracts, and mark every blank as an assumption someone must confirm.