Software vendors built their business models on a reliable assumption: renewal is the default, and the work of selling happens once. That assumption is weakening, and the reasons are structural rather than cyclical.
Why renewal used to be automatic
The economics of subscription software depend on retention. Acquiring a customer costs a great deal — sales commission, implementation, onboarding — and the investment is recovered over subsequent years. A vendor that acquires a customer and keeps them for eight years has an excellent business. The same vendor with a three-year average life may not have one at all.
Retention was historically high because switching was genuinely hard. Data lived in the vendor’s format, staff were trained on the interface, and integrations had been built against its interfaces. The cost of leaving frequently exceeded several years of subscription fees, so buyers renewed regardless of whether they were satisfied.
That is not primarily a statement about product quality. It is a statement about switching costs.
What changed
Seat counts became visible. Usage analytics are now standard, and finance departments have learned to run them before renewal. A contract for 500 seats with 280 active users is a straightforward saving, and it requires no migration and no risk. Reducing seats is the easiest cost reduction available in a software budget.
Procurement got professional. Software purchasing that once sat with departmental budget holders now routinely passes through a procurement function whose job is explicitly to reduce spend and which has benchmark pricing across comparable organisations. Vendors that priced opportunistically are discovering that their customers now compare notes.
Consolidation pressure. Most organisations accumulated overlapping tools through years of departmental purchasing. Rationalising to platforms already paid for eliminates spend without losing capability — and the incumbent platform vendor is happy to encourage it.
Data portability improved. Partly through regulation, partly through competitive pressure, export is more feasible than it was. The moat has not disappeared, but it is shallower.
The metrics that reveal it
Software companies report net revenue retention — revenue from existing customers this year against the same cohort last year, including expansion, contraction and churn. Above 100 percent means the existing base grew without new customers.
The figure is genuinely informative, and it is also the metric most vulnerable to presentation choices.
- Which customers count. Cohorts frequently exclude the smallest customers, where churn is highest. The reported figure then describes a subset chosen partly because it performs well.
- Gross versus net. Net retention nets expansion against churn. A vendor losing 20 percent of customers while expanding heavily within the rest can report healthy net retention while its customer base erodes. Gross retention — which excludes expansion — is the more revealing number and is disclosed less often.
- Timing. Multi-year contracts defer the moment of truth. A vendor that shifted to three-year terms will show strong retention for two years regardless of satisfaction.
The leading indicator ahead of any of these is billings or remaining performance obligations rather than recognised revenue. Subscription accounting recognises revenue across the contract term, so reported revenue reflects decisions made quarters ago. Deteriorating renewals appear in new bookings well before they appear in the income statement.
The buyer’s position
For an organisation on the purchasing side, the negotiating position is better than it has been in some time, and there are a few things worth knowing.
- Measure utilisation before the renewal conversation, not during it. Actual active users is the single most useful fact in the negotiation and vendors do not volunteer it.
- Start early. A renewal discussed six months out is negotiable. One discussed three weeks before expiry is not, and vendors structure their calendars accordingly.
- Auto-renewal and uplift clauses. Many contracts renew automatically with a built-in annual increase unless notice is given by a specified date. That date is usually well before expiry, and missing it forfeits the negotiation entirely.
- Quarter end and fiscal year end. Vendor sales teams have quota deadlines, and discount authority expands as those approach. This is the most reliable timing advantage available to a buyer.
- Price the exit before you need it. Knowing what migration would actually cost — in time, integration work and disruption — converts a vague sense of being trapped into a number you can negotiate against.
That last point is the one most organisations skip, and it is the one that determines everything else. A buyer who has not costed the alternative has no position, and both sides know it. It is the same calculation that governs infrastructure decisions more broadly — we worked through a version of it in the economics of cloud repatriation.
