Sales Bonuses Tied to Invoice Value: The Silent Margin Killer

The single most expensive structural decision I made in my first three years running an agency was paying our sales team on invoice value rather than on real margin. I made the decision because everyone else did. I unmade it about four years later, after watching the same pattern of misaligned incentive play out on three of our biggest engagements in a row.

This piece is the case for switching the bonus structure, the math underneath it, and the conversation to have with the sales team when you do. It's the conversation I wish someone had laid out for me before I had to figure it out the expensive way.

What goes wrong with invoice-based bonuses

The standard agency sales compensation: base salary plus a commission that's a percentage of the invoiced revenue on engagements the salesperson closes. Typical rates in the UK and US: 3–8% of first-year revenue, sometimes with a tail on year 2 of an ongoing engagement.

The structure feels obvious. Sales brings in revenue; sales gets a share. The challenge: it incentivises three behaviours that, on average, hurt agency margin.

Behaviour 1: pushing estimates up rather than estimating accurately

A salesperson paid on invoice value has a direct interest in the number on the estimate being as high as the client will accept. The cleanest way to push it up is to add work that the client side wasn't initially asking for ("while we're here, you should also consider rebuilding the auth flow") or to inflate role hours ("we'll need 80 hours of architect time for this, not 50").

The first push, where it lands, often produces good extra work. The second push compresses delivery against an estimate that the team can't honestly hit, and the project lands at lower margin than it should have.

We did the analysis on our own engagements over an eighteen-month window. Estimates that came in 15% or more above the rough-order-of-magnitude that the project lead had ballparked, after the salesperson had been involved in pricing, landed on average 9 percentage points lower in real margin than the on-ROM estimates. The salesperson was paid more on the high-priced estimate; the agency made less.

Behaviour 2: accepting role mixes that under-budget senior time

A salesperson trying to keep an estimate at the price the client wants, while still hitting the right invoice value, has a lever: under-budget the senior time and let the team absorb it during delivery.

The estimate goes out at, say, "200 hours of mid-level engineering" instead of "100 hours of mid + 50 hours of senior". The price stays attractive; the client signs; the team delivers. During delivery the senior absorbs the 50 hours of work that wasn't budgeted (because the work needed senior judgement and the mids couldn't crack it), and the labour-margin report shows the project hitting its hours. The real margin, computed against employee cost, is 15 percentage points lower than the report claims.

The salesperson is paid against the original invoice. The agency absorbs the variance.

Behaviour 3: closing engagements where scope is likely to grow

The third behaviour is the most subtle. A salesperson paid on first-year revenue has an interest in landing engagements that will grow. Engagements where the client has signalled, in informal conversation, that they're likely to want more work after the initial scope.

This sounds positive — growing engagements are good for the agency. The problem: scope-growing engagements are also the ones most likely to have unbilled scope creep, because the same client who's signalling future expansion is signalling that they expect a flexible relationship now. The salesperson who landed the engagement is now incentivised to keep that flexibility going (no change orders, friendly atmosphere, soft scope boundaries) because the next engagement depends on the relationship.

The team delivers more than was scoped, doesn't push for change orders, lands the project at compressed margin. The salesperson, meanwhile, lands the next engagement and gets paid on the new invoice.

The agency is, in effect, paying its sales team to maintain client relationships at the cost of project margin. That's not what the bonus was meant to fund.

The math: what invoice-based bonuses actually pay for

Take a worked example. Sales has closed two engagements this quarter:

• Engagement A: £140k invoice, real margin 32% (£44.8k of margin)
• Engagement B: £180k invoice, real margin 11% (£19.8k of margin)

At a 5% commission on invoice value, the salesperson earns:

• A: £7,000
• B: £9,000
• Total: £16,000

The salesperson has been paid 56% more on Engagement B (£9,000) than on A (£7,000), even though Engagement A produced more than twice the real margin (£44.8k vs £19.8k).

The agency, meanwhile, takes home:

• A: £44.8k − £7k bonus = £37.8k
• B: £19.8k − £9k bonus = £10.8k

The pre-bonus margin gap (£44.8k vs £19.8k = 2.3× difference) becomes a post-bonus margin gap (£37.8k vs £10.8k = 3.5× difference). The bonus structure has amplified the margin difference, paying disproportionately on the worse engagement.

Multiply this across a year of sales activity and the agency is paying the sales team to bring in lower-margin work. Not because anyone is dishonest — because the bonus structure points the team that direction.

The fix: bonuses on real margin

The structural fix is to pay the bonus on the project's real margin (its saldo), rather than on invoiced revenue. The same total bonus pool, allocated by margin produced rather than revenue closed.

Take the same example. Switching to a 25% bonus on real margin:

• A: 25% of £44.8k = £11,200
• B: 25% of £19.8k = £4,950
• Total: £16,150 (essentially the same as the £16,000 under invoice-based)

The aggregate compensation has barely moved. The allocation has flipped. The salesperson now earns 60% more on Engagement A than on B, which matches the relative margin contribution.

What this changes about salesperson behaviour:

• Estimates get more honest, not bigger. A bigger estimate that compresses margin is now bad for the bonus, not good. The salesperson has every incentive to defend an accurate estimate.
• Role mixes get defended against under-budgeting. Under-budgeting senior time produces a project that lands at compressed margin, which now hits the bonus directly.
• Scope creep gets pushed back on. The engagement that grows without change orders is the engagement whose margin compresses, which now hits the salesperson's pocket.
• Pricing reviews happen at the salesperson's request. A retainer that's being eaten by hour creep is now actively bad for the salesperson's bonus, and they have an incentive to ask leadership to reprice it.

The behaviour change, where we've watched this install, has been visible within one quarter. Salespeople start asking, in pre-pitch internal reviews, "is this engagement going to land at margin?" — a question they had no reason to ask under the old structure.

The conversation with the sales team

The hard part of installing the new bonus structure is not the math. The math is roughly cost-neutral over a year. The hard part is the conversation with the sales team, because the new structure removes a perceived guarantee.

Under invoice-based bonuses, the salesperson knows roughly what they'll earn the day they close the deal. The invoice value is set; the bonus is a fixed percentage; the timing is predictable. Under margin-based bonuses, the bonus is contingent on a number that depends on how the project is delivered — which the salesperson doesn't fully control.

Three structural elements make the conversation work:

Element 1: real-time visibility on the salesperson's bonus

The salesperson has to be able to see, on a dashboard, what the running margin is on each engagement they've closed. This is not just for transparency — it's for action. A salesperson who sees that one of their engagements is sliding into margin compression has an incentive to flag it to leadership early, which is good for everyone.

This is one of the things that requires a continuous P&L. With a monthly closing pack, the salesperson sees the bonus three weeks after the project's behaviour changed, which is too late to act. With a continuous P&L, the bonus is a number that updates daily.

Element 2: a base salary that doesn't depend on margin

The salesperson's base pay should be high enough to live on without the variable component. The variable component is what changes behaviour; the base is what makes the change tolerable. We aim for a 70/30 base/variable split, where most salespeople land in the 60/40 range.

If the base salary has been low because the variable was assumed to be substantial, the switch to margin-based bonuses requires a base salary increase. Account for this. The aggregate compensation over a year is roughly the same; the timing is different.

Element 3: a clear definition of "real margin" that the salesperson can audit

The margin number has to be defensible to the person whose bonus depends on it. That means:

• The cost-side numbers (employee hourly costs, overhead allocation, sub-project membership) have to be consistent and auditable.
• The salesperson can see how their bonus number was computed for each engagement.
• Disputes about the margin calculation have a clear path: a meeting with finance and operations where the numbers are walked through.

We've found that within the first quarter, the salespeople become some of the most knowledgeable people about the agency's project economics, because they have a personal stake in understanding the calculation. That's a positive externality of the structure that we didn't predict.

What we've seen install in practice

Three observations from agencies that have switched, including ours.

• The salespeople who were already aligned with the agency's margin reality are unaffected. Their bonuses come out roughly the same. They're often the most enthusiastic supporters of the change, because it formalises what they were already trying to do.
• The salespeople who were optimising for invoice value see their bonuses compress for the first quarter or two, then recover as their behaviour shifts. A 10–20% temporary compression is typical. By the third quarter, most are at or above their previous numbers.
• About 10% of salespeople don't make the transition. Either they leave (preferring to find an agency with the old structure), or they're managed out because their pipeline persistently produces low-margin engagements. This is uncomfortable but, in retrospect, necessary — the agency was paying them to bring in work that didn't pay for itself.

The aggregate effect on agency margin in the first year, in our data: 5–8 percentage points of recovered margin, almost all of it from the behaviour change rather than from the bonus pool itself.

What this requires structurally

For the bonus structure to work, the underlying systems have to deliver three things:

• A real margin number that's continuously available. The closing-pack approach to margin reporting cannot support a margin-based bonus, because the number arrives weeks after the work it represents. A live P&L is structurally required.
• A clear audit trail of cost-side data. Hourly costs, role rates, overhead allocations — all of these have to be defensible if a bonus is going to be paid against the resulting margin. Salaries change; the audit trail has to record the cost number that applied at the time the worklog was logged.
• A discipline of separating sub-projects. The build's margin and the support retainer's margin have to be reportable separately, so a bonus paid on the build is computed against the build's actuals — not against a Frankenstein number that includes the post-launch support.

These are the same three things that justify a financial layer like Saldo. I didn't build Saldo for sales bonuses, but the bonus restructure was one of the first things the agency I was CTO of did once the layer was in place, and it was one of the highest-impact changes of the first year on the system.

Where this lands

Most agencies will not change their bonus structure tomorrow. The status quo is sticky, the conversation with the sales team is uncomfortable, and the systems work to support a margin-based bonus has to be in place before the change can be made credibly.

The point of this article is not that you should make the change next quarter. The point is that the structure you have right now, if it's based on invoice value, is silently working against your margin every quarter you keep it. The longer it stays in place, the more behaviour reinforces the misalignment.

If you want to see what the migration looks like — the live margin number, the audit trail, the bonus-relevant calculation — that's part of what Saldo provides. Pricing is on the pricing page; the reference customer's two-year experience includes the bonus restructure as one of the early-quarter changes.

The shorter version: paying sales on invoice rewards landing big estimates. Paying sales on margin rewards landing profitable engagements. The two are not the same business, and the bonus structure is the most direct way to choose which business you're in.