Two things happened to collaboration budgets this year. Artificial intelligence add-ons arrived at roughly thirty dollars per user per month, and finance functions that had tolerated tool sprawl for a decade suddenly wanted the list. Those are the same event. The add-on is what made the base spend visible.
Nobody is cutting collaboration tools to save money on collaboration tools. They are cutting them to pay for the assistant everyone has already promised
Understanding that is the difference between a rationalisation that sticks and one that quietly reverses in six months.
What the inventory usually finds
A typical mid-sized group with a thousand staff runs a primary suite, a chat platform, a separate video product, two or three project tools bought by different departments, a design tool, a whiteboard, a documentation product, a form builder, an electronic signature service, at least two file-sharing arrangements and a scattering of departmental subscriptions paid on cards. The headline number is rarely the shock. The shock is the overlap: three products that do task tracking, two that do documents, and a video capability bought twice because the suite's version was judged inadequate in 2021 and nobody revisited it.
The four categories of waste, in order of how easy they are to remove
Unused licences. People who left, people who never activated, seats bought for a project that ended. Pure savings, no behavioural change, usually recoverable at the next renewal rather than immediately. Duplicate capability. The same job done by two products. Real savings, moderate disruption, and the negotiation is internal rather than with the vendor. Over-tiering. Everyone on the premium plan because it was simpler, when a third of users need nothing beyond the base. Easy to model, tedious to administer, and frequently worth more than the duplicate-tool exercise. Genuinely redundant products. Tools that a small group loves and the organisation does not need. This is where consolidation programmes die, because the affected team is always articulate and the saving is always modest.
| Category | What the review must establish |
|---|---|
| Unused licences | Which seats are unused and when contract terms allow removal |
| Duplicate capability | Whether shared work needs two products and what a move would disrupt |
| Over-tiering | Which roles need premium capabilities rather than the base plan |
| Redundant products | Whether a specialist need is real and whether removal costs more than it saves |
Qualitative summary of this article's source text, not a measured outcome or performance estimate.
Consolidation has costs the business case usually omits
Migration effort, which is mostly people's time and is never zero. Content left behind, because export rarely preserves structure and nobody checks until someone needs the old thing. Retraining, which is small per person and large in aggregate. Lost capability, where the replacement does eighty per cent of the job and the missing twenty per cent belongs to the team that complains loudest. And concentration risk, which is the one nobody prices until a single supplier has an outage. A credible case names all five and still shows a saving. A case that shows only licence cost will be approved and will then disappoint.
The rule that makes the decision tractable
Consolidate where the work is shared and allow specialisation where it is not. Communication, documents, file storage and the primary task system should be single and universal, because their value comes from everyone being in the same place. Specialist tooling — design, engineering, analytics, creative production — should be left alone unless the licence cost is genuinely material, because forcing a specialist onto a general tool destroys more value than the subscription costs. Most failed consolidations violate the second half of that rule while congratulating themselves on the first.
What to do with the assistant add-on
Do not buy it for everyone in the first year. Buy it for the roles where drafting, summarising and searching are a large share of the day, measure whether anything changed, and expand on evidence. Vendors will price an all-staff commitment attractively and the discount is real; the utilisation is usually not, and a year of low utilisation is a weak position from which to renew.
Practical Guidance for SaaS Spend Review
- List every tool including card-paid subscriptions, then map capability overlap.
- Pull actual activity data, not seat counts, from each administrative console.
- Right-size tiers before removing products; it is faster and less political.
- Price migration, retraining and lost capability into the case.
- Consolidate shared workflows, protect specialist tooling.
- Align renewal dates so future decisions are made together.
- Buy assistant seats by role, then expand on measured use.
- Name an owner for the application estate with authority over purchases.
The Regional Angle
Start with the structure of the organisation, because regional groups make this harder than the headcount suggests. A holding company with operating subsidiaries across several countries, free zone and mainland entities, and joint ventures with partners who have their own systems does not have one collaboration estate; it has six, each procured locally, several contracted in different currencies, and at least one that cannot be consolidated because the joint venture agreement gives the partner a say. Before building a consolidation plan, establish which entities can actually be brought onto a single tenancy — legally, contractually and in terms of who owns the data — and treat the rest as a federation to be standardised rather than merged. A programme that assumes a single tenancy and discovers the joint venture halfway through loses a year. The second is the channel that does not appear in any procurement inventory and carries much of the actual work. Customer, supplier and internal coordination in this region runs substantially through consumer messaging apps, and that usage is free, invisible to finance, and therefore never counted as a collaboration tool. Any consolidation exercise that reduces the sanctioned toolset without addressing that channel simply pushes more traffic into it, and the records go with the traffic. Decide deliberately what the messaging channel is for, provide a sanctioned business equivalent where records matter — orders, approvals, anything a dispute would turn on — and accept the informal channel for the rest. The savings from removing a project tool are trivial next to the cost of a commercial dispute argued from a personal phone. The third is a contracting detail that changes the arithmetic. Regional entities frequently buy software through local resellers rather than direct, in local currency, on annual terms with automatic renewal and narrow cancellation windows, sometimes bundled with implementation or support services that make the licence line hard to isolate. Consolidation savings that assume a mid-term exit usually evaporate on contact with those terms. Read the notice periods first, build the plan around the renewal calendar rather than the fiscal year, and use the reseller relationship deliberately — they can often restructure tiers and align dates in a way that direct terms will not, provided you ask before the window closes rather than after.
The objection worth taking seriously
The strongest objection is that consolidation programmes are a tax on the people doing the work, dressed up as financial discipline. The savings are modest — a few hundred dollars per head annually in most estates — while the disruption falls on teams who chose their tools because the standard option did not fit. Everyone has seen a rationalisation that removed a product a small team depended on, replaced it with something worse, and quietly tolerated a workaround that cost more in lost time than the licence ever did. And the sprawl frequently reflects a real failure of the central platform rather than indiscipline at the edges. That criticism is fair and it should shape how the exercise is run. The part that holds is narrower than the usual programme and much less contentious. Unused licences and wrong tiers are pure waste with no behavioural consequence at all, and in most estates they account for the majority of the recoverable spend — which means the defensible savings can be taken without asking anybody to change tools. Beyond that, the honest question is not what a product costs but whether the work it holds needs to be visible to people outside the team that bought it. Where the answer is yes, consolidate. Where it is no, leave it alone and stop counting it as a problem. Run the exercise that way and it produces a saving and a shorter list. Run it as a uniformity project and it produces a shorter list, a slower organisation, and a shadow estate within the year.
Common Questions
How much can we realistically save?
In a typical estate, somewhere between fifteen and thirty per cent of collaboration licence spend, with most of it from unused seats and over-tiering rather than from removing products.
Should we consolidate before or after buying artificial intelligence add-ons?
Before, if you can. The add-on is priced per seat, so removing unused seats and wrong tiers first reduces the cost of everything that follows.
What if a team refuses to move?
Ask whether the work in that tool needs to be visible outside the team. If it does, the move is not optional. If it does not, the exception is cheap and you should grant it explicitly rather than fight it.
What should we expect over the next twelve months?
Expect suite vendors to keep bundling capability that standalone products charge for, which strengthens the consolidation case each year without you doing anything. Expect assistant pricing to fragment into tiers and consumption models as all-seat adoption disappoints. Expect finance to ask for usage evidence at renewal rather than accepting seat counts. And expect at least one of your specialist tools to be acquired by a larger vendor, which will make the decision for you.
SaaS Spend Review — we find the unused seats and wrong tiers first, then tell you which products are genuinely worth consolidating.