Published: 21 May 2026 · Roy Morken, Datafolka

A four-week SaaS audit: practical template for cutting 20–40 % of licence costs

A manufacturing company with 180 employees was spending NOK 612,000 a year on SaaS. Three tools did almost the same thing: one Norwegian project tool, one international one brought in through an acquisition, and a free tier that a department head had scaled up on a personal card. After the audit, the annual cost was NOK 361,000. A 41 percent cut. No functionality lost.

This is the case Roy Morken, co-founder of Datafolka, elaborated on in Digi on 20 May under the title Too many companies are overpaying for their SaaS stack . The op-ed format allowed room to show the problem and three key actions, but not the execution template itself. That is what follows here.

The rest of this article is a concrete four-week plan: a spreadsheet template for the register, a usage-data checklist per vendor, negotiation scripts in three variants, and a calculation that shows when "all-in-one" actually pays off and when it ends up costing more.

Why the stack grows without anyone noticing

In the audits we have been involved in, the same three patterns recur.

First: decentralised purchasing. Marketing takes HubSpot, sales adds Salesforce, support buys Intercom. Each decision is rational in isolation. Nobody sees the full picture until the accountant asks about a cost code she does not recognise.

Then: the free tier that grows. Notion or Figma starts with three users and a personal card. Three months later there are 40 people in the workspace and someone has upgraded to the Business plan "because it was needed". The invoice ends up on the wrong cost centre.

Finally: licence spillover. A developer leaves, but the account sits active for two years because nobody cross-references the offboarding list against the SaaS register. That register does not exist.

That is where we begin.

Phase 1: Build the register (week 1)

A SaaS register is a spreadsheet. Not a tool. A spreadsheet is what comes out of this phase, and it is the most important artefact in the entire audit.

The template we use has ten columns:

Data comes from three places.

Invoice pull from accounting. Export all vendors from the last 12 months and filter for anything that looks like SaaS. This gives you between 80 and 90 percent of the stack.

SSO logs. If you have Okta, Azure AD, Google Workspace, or another identity provider, export the list of applications and logins for the last 90 days. This is where you find what invoices miss: tools billed via a marketplace aggregator or a third-party vendor.

Ask department heads directly. Send one email: "Which tools does your team use that cost money?" Draw a line between what they report and what invoices show. The gap is almost always instructive.

A realistic first week ends with 30 to 50 entries. We have seen 90 at an organisation with 220 employees. It should be uncomfortable to read.

Phase 2: Extract usage data (week 2)

The licence count in the register is what the vendor invoices you for. The usage count is what you actually get in return. The gap is the cutting potential.

Most major vendors expose usage data, but not all make it obvious. Here is the checklist we go through:

Microsoft 365 / Entra ID. Admin Center > Reports > Usage. Export the "Active users" reports per product (Teams, OneDrive, SharePoint, Exchange). Users with zero activity in the last 30 days are candidates for downgrading from E3 to E1, or for removing the licence entirely.

Google Workspace. Admin Console > Reports > Apps Usage. Same principle. Inactive Gmail accounts are often former employees who were never removed. Every licence you are not paying for is pure savings.

Slack. Customize Workspace > Analytics. Export the member list. Filter by "Last active". Workspaces with more than 50 inactive users should be downgraded to Pro or moved to a consolidated workspace.

GitHub. Organization settings > Billing > Plans. Cross-reference the seat list with git activity in the last 90 days. Users with zero commits and zero PRs in the last 90 days are candidates for removal.

Atlassian (Jira, Confluence). User Management > User Access. Export active licences. Confluence is particularly prone to spillover, because the licence often follows Jira access.

Salesforce. Setup > Users > Login History. Users with no login in the last 60 days are clear candidates. Salesforce is also one of the few tools where edition downgrading (Enterprise to Professional) yields real savings if you do not use specific Enterprise features.

HubSpot. Settings > Account > Users & Teams. Seats are split per hub (Marketing, Sales, Service). Export activity per hub. It is not unusual for an organisation to be paying for Sales Hub Professional for 12 users when only 4 actually use it.

One tip: when the tool's dashboard does not give you what you need, ask the account manager for a usage report directly. We have obtained data from Salesforce and Atlassian that the dashboard does not expose at all. They have it. Ask.

Mark each entry in the register with one of three statuses: keep, downgrade, or phase out.

Phase 3: Negotiate the renewal (week 3)

This is the simplest lever in the entire audit. It is also the one people resist most.

Send an email 60 days before renewal. Not 30, not 14. Sixty days before gives the vendor time to escalate internally, and it gives you time to actually switch if the conversation goes nowhere.

Three templates we typically use. Adapt the tone to your own voice, but the structure is deliberate.

Version 1: Reducing licence count

Hi [Account Manager],

We are approaching the renewal of [tool] on 1 November.

We have reviewed our usage over the past three months.
Of 42 licences, 27 have been active in the last 30 days.
The rest are former employees or teams that no longer use
the tool.

For the renewal, we would like to reduce from 42 to 30 seats.

Can you confirm that is possible, and whether it triggers any
changes to the per-seat price or contract terms?

Kind regards,
[Name]

Expected response: yes, and sometimes the vendor offers to keep the licence count in exchange for a 10–15 percent discount. Both outcomes are good.

Version 2: Switching to a cheaper plan

Hi [Account Manager],

For the renewal on 1 November, we are considering moving
from the [Enterprise/Business] plan to [Pro/Standard].

We have reviewed our feature usage and find that we do not
actively use [specific Enterprise feature, e.g.
SSO, advanced analytics, dedicated support]. Of these,
only [one specific feature] is something we need — and that
is available on the [cheaper plan] as well.

What is the price difference, and is there anything we should
be aware of when making the switch?

Kind regards,
[Name]

The trick here is to be specific about which Enterprise features you are not using. Vague statements ("we don't use everything") get vague responses. Specific points get specific offers.

Version 3: Threat to switch, with a named alternative

Hi [Account Manager],

We are approaching renewal on 1 November and have evaluated
the market.

[Competitor] offers [feature parity or better]
at [specific percentage] lower annual cost for the same
number of seats. We have tested in sandbox over the past
two weeks, and the transition is manageable.

We wanted to give you the opportunity to come back with
a proposal before we make the switch. What can you offer?

Kind regards,
[Name]

This only works if you have actually tested the alternative and are genuinely willing to switch. Bluffing gets seen through. In audits we have been involved in, this kind of email yields 15–30 percent discount when it is real. It yields zero when it is a bluff, because the account manager calls a colleague and finds out that the sandbox was never logged into.

Once you have sent the email, the vendor will suggest a call. Take it. But first send a written reply: "Yes, I can take a call on Tuesday at 10. To make it efficient, could you have a concrete offer ready on [the point above]?" That way the call gets to the point.

Phase 4: Consider consolidation, but calculate first (week 4)

The final lever is consolidation: replacing three tools with one that covers most of the same ground. It always sounds appealing. It is not always right.

Here is the calculation from a consolidation we walked a customer through. They were considering switching from Asana + Notion + Loom to an "all-in-one" platform covering project management, documents, and video messages.

All-in-one was NOK 138,000 more expensive in year 1. Even after removing migration costs in year 2, the ongoing cost was still NOK 198,000 against NOK 204,000. Barely any savings, against the considerable risk of putting everything with a single vendor.

The decision was to keep three tools, but downgrade Asana from 40 to 30 seats (12 inactive), and switch Notion Business to Plus for 60 users (they were not using the audit log). Total saving: NOK 38,000 per year. Zero migration work.

Consolidation pays off when one of three things is true: the existing tools are so underused that 70 percent of licences can be dropped, you have a real integration problem costing developer hours every month, or the all-in-one vendor is already deep in your stack and is adding other modules at near-zero cost.

If none of those apply, calculate first. Migration hours are what most often eats the saving.

The most important action

After all the audits we have been involved in, one finding surprises us least: organisations that say "we'll do it next quarter" never do. Organisations that set a date this month and start filling in the register on a Monday land a cut of 20–40 percent within eight weeks.

It is not a tooling problem. It is a decision problem.

If you have a renewal due within 90 days, start Phase 1 this week. If you do not have internal capacity to run this process but want a thinking partner: we take on this kind of audit. Send an email to Roy.Morken@Datafolka.no and we will set up a call.

Roy Morken, co-founder of Datafolka, elaborated on the background for this method in Digi on 20 May: Too many companies are overpaying for their SaaS stack .