A five-year contract is not five years of revenue

Veterinary medicine in the United States has been consolidating for more than a decade. Groups backed by private capital have bought thousands of independent practices, and for a software company like ours they are now among the largest customers. They negotiate differently from a single hospital. A typical agreement covers dozens or hundreds of hospitals, runs for three to five years, starts at a discount that rises over time, includes credits for implementation and data migration, and often allows the group to end the agreement early.

These agreements are good business. They bring hundreds of hospitals onto a platform at once, and the groups are demanding customers whose feedback improves the product. But they have also changed how software companies describe their growth, in a way that I think misleads boards more than it misleads investors.

What the signature actually commits

When a large agreement is signed, the instinct is to announce its total contract value: the sum of every payment due over the full term. In 2023 we signed a five-year agreement with Bramblecote Veterinary Partners, a consolidator, covering 210 hospitals. Its total value on the cover page was $38 million. Bramblecote could also terminate for convenience after 24 months, with 90 days' notice and a modest fee.

Under ASC 606, the US revenue standard, the contract term for accounting purposes is the period during which both parties have enforceable rights and obligations. A customer's right to walk away without a substantive penalty shortens that period. For revenue purposes, our Bramblecote agreement was in substance a two-year contract with the possibility of more, not a five-year one. I am describing the principle in general terms; the answer for any particular contract depends on its wording and its penalties, and we worked through ours with our auditors.

The difference reaches the disclosures. Listed companies report remaining performance obligations, the contracted revenue not yet recognised. If a company's formal disclosure correctly excludes cancellable years while its sales metrics and its commentary include them, two versions of backlog begin to circulate. Investors usually find the formal one in the filings. Boards often see the other, because it is the one in the sales review.

Ramps, credits and the price we really agreed

The second complication is price. Large agreements rarely charge the same amount each year. Ours with Bramblecote charged 70 per cent of list price in the first year and 85 per cent in the second, reaching full price from year three, with $2.2 million of implementation credits in the first year. Averaged across the cover page's five years, the price per hospital looked about 9 per cent below list. Within the enforceable two years, once the credits were counted, it was about 36 per cent below.

This matters because ramps and credits push the profitable years of a contract into the period the customer may never reach. A five-year agreement priced to earn most of its margin in years three to five is, in practice, a two-year agreement at a steep discount plus an option for us to earn a fuller price if the customer stays. That option has value. It is not revenue, and presenting it as if it were misleads the people who plan hiring and spending against it.

I now show our board a schedule for every agreement above $1 million of annual value: the enforceable term, the revenue within it, the effective discount within it, and the revenue that depends on renewal or on the customer choosing not to terminate. The enforceable total for all such agreements is much smaller than their headline values would suggest. It is also the number we can defend to an auditor and plan cash against.

Pay for what the customer cannot cancel

The deepest problem is not accounting but incentive. Before 2023 our enterprise sales team, like many in software, was paid commission partly on total contract value. A representative who added a fifth year to an agreement, even one the customer could cancel after the second, earned more. So did a representative who traded a deeper first-year discount for an extra year. The plan rewarded exactly the terms that weakened the enforceable contract, and the people negotiating were acting rationally on the plan we had given them.

We changed it. Commission is now paid on annual contract value within the enforceable term, with a further payment when a customer renews or passes the point at which it could have terminated. Termination rights, price ramps beyond an agreed limit and credits above a set level need approval from a deal desk that includes finance and legal. Our head of enterprise sales argued that the change would make us less competitive against rivals still paying on total value, and in two large bids it may well have done so. I accept that cost.

The result is a contracted backlog that is smaller than some peers would report for a business of our size, and one that I trust. When a large group renews, the renewal shows up as genuinely new commitment rather than as a year we had already counted. Bramblecote passed its termination point in 2025 and stayed, and that was the first moment its later years entered our backlog.

There is a fair objection: customers rarely exercise termination rights, so excluding the cancellable years understates what will probably be collected. Our own history supports part of that. Most large groups do stay. But "probably" is a forecast, and a forecast belongs in a forecast, labelled as one, not in a figure presented as contracted. We track how often groups actually exercise their termination rights and publish that rate beside the enforceable figure, so readers can judge the gap for themselves.

Directors do not need to master revenue recognition to ask the right questions. For any large multi-year agreement, ask how long the customer is actually bound, what it pays within that period, and what the sales team earned on it. If the answers differ from the press release, the board is reading a sales announcement rather than a commitment. Those three answers take a finance team an afternoon to assemble, and in my experience they change a board's conversation about a quarter's bookings more than any new metric does.