Revenue Optimization: Where the Money Actually Leaks

Almost every revenue optimization programme is an acquisition programme wearing a broader word. More leads, more meetings, more pipeline, more reps. It is not that this is wrong — it is that it is one of four places revenue leaks, and reliably the most expensive one to plug.

The reason is not that anybody has analysed it and concluded acquisition is the best lever. It is that acquisition is the only leak with a department, a budget line and an obvious owner. The other three leak quietly, across functions, in amounts nobody has added up.

This article adds them up.

The four leaks

Revenue is a chain, and it leaks at every joint:

1. Acquisition — you do not reach, or do not qualify, enough of the right people. Opportunities that never existed.
2. Conversion — opportunities enter the pipeline and do not finish. Deals you had and did not close.
3. Realisation — you win, and you are not paid what the deal was worth. Discount, scope creep, unbilled work, bad packaging.
4. Retention and expansion — you keep the customer or you do not, and you grow them or you do not.

Most companies instrument the first two well, the third badly, and the fourth only in aggregate. Which turns out to be exactly backwards from where the money is.

The arithmetic that reverses the usual advice

Here is a model. The numbers are invented and labelled as such; run it with yours, which will take an afternoon.

Opportunities created per year	1,000
Win rate	20%
List price (annual)	$20,000
Average discount	15%
Realised price	$17,000
New business	$3.4M
Installed base entering the year	$10.0M
Gross retention	85%
Expansion	8%
Net revenue retention	93%
Base contributes	$9.3M
Total	$12.7M

Now improve each leak by ten percent, relative. Not ten points — ten percent of whatever you currently have, which is roughly what a good year of focused work buys at any of them.

Leak	What changes	New business	Gain
1. Acquisition	1,000 → 1,100 opportunities	$3.74M	$340k
2. Conversion	Win rate 20% → 22%	$3.74M	$340k
3. Realisation	Discount 15% → 6.5%	$3.74M	$340k

Identical. All three, to the dollar.

That is not a coincidence in the model, it is a property of the structure: new business is a product of those three terms, so ten percent applied to any factor moves the product by exactly ten percent. Which produces the sentence this article exists for:

In a multiplicative chain, the question is never where the biggest gain is. Every gain is the same size. The question is where ten percent is cheapest to buy, and how long it lasts once you have it.

And then the fourth leak, which breaks the symmetry:

Leak	What changes	Base contributes	Gain
4. Retention	NRR 93% → 102.3%	$10.23M	$930k

Nearly three times the others — 2.7× — for the same relative improvement, and for one structural reason: it applies to the installed base rather than to the new business, and the installed base is the larger number. In any company past its earliest years, it is much the larger number.

Then it recurs. Next year's base includes this year's new business, and the improved retention rate applies to that too. Acquisition gains have to be bought again every January; retention gains compound.

So: cost and durability, not size

If the gains are equal, ranking them by size is meaningless. Rank them by what a point costs and whether it stays bought.

Acquisition. Costs money, directly — spend, headcount, or both — in the most competitive market you participate in, because every competitor is bidding for the same attention. It is also the least durable: stop paying and the gain stops the same quarter. And it is diluted by everything downstream, since each new opportunity must still survive your conversion, your discounting and your churn before it becomes revenue. It is the only leak where the fix is purchased rather than changed, which is precisely why it is reached for first — a purchase has an owner and an invoice.

Worth noting what that purchase really buys: more hours pointed at buyers. That is a capacity question, and the arithmetic of capacity says most teams have a large unused pool of it already on the payroll before they buy more.

Conversion. Costs coaching and enablement time rather than money, and is partially durable — skills persist, though people leave and take them. It is also the leak where diagnosis matters most, because "improve win rate" is not an instruction; the stage-by-stage decomposition is what turns it into one.

Realisation. Frequently costs nothing but discipline, which makes it the cheapest point on the board and the most consistently ignored. Discount approval thresholds that are actually enforced. A price floor. Not conceding in the last week of the quarter because the forecast needs it. The reason it gets skipped is that every individual discount has a good story attached, and the leak is only visible in aggregate — which is also why the fix has to be structural rather than a request to hold firmer.

It helps enormously that most price objections are not requests for a smaller number at all, and treating them as such is the actual leak.

Retention and expansion. The slowest to move and the only one that compounds. It usually costs product and service investment rather than sales investment, which is why a sales-led revenue programme tends to skip it — the work is not in the department running the programme. It is nevertheless the largest prize in the model above by a factor of nearly three.

Sequencing

A workable order, assuming a base large enough for the arithmetic above to hold:

1. Size all four in your own numbers. An afternoon, and it needs no more than you already have — the same discipline as deriving coverage from your own win rate rather than a published multiple. Until this is done, every argument about where to invest is a matter of temperament.
2. Take realisation first. It is usually the cheapest point available and it requires no new spend, only enforcement. It is also the fastest to show up.
3. Start retention immediately, in parallel, expecting nothing for a year. It is slow, it compounds, and starting it late is the single most expensive scheduling decision in this article. Anything begun here pays every subsequent year.
4. Then conversion, once you can name which stage is leaking rather than pointing at the aggregate.
5. Then acquisition — into better ratios, so the opportunities you buy are worth more than the ones you were buying before.

The ordering principle is the same one that governs capacity and hiring: fix what multiplies before you buy more of what gets multiplied.

What to instrument

Most of this cannot be argued about productively until four numbers exist. Three are common; one almost never is.

Opportunities created and stage-to-stage conversion. Standard, and covered in the metrics that change a decision.

Realised price against list, by rep and by segment. Rare, and the single highest-value instrumentation on this list because it makes an invisible leak visible. It also diagnoses itself: uniform discounting across every rep means the price is wrong, concentrated discounting means a skills gap.

Gross retention and net retention, reported separately. This is the one that matters most and gets fudged most often, because reporting only NRR hides churn behind expansion.

And a fourth worth adding: churn reasons, recorded at the point of churn, in categories a human chose rather than a dropdown default. Most retention programmes fail because they are designed against a guess about why customers leave, and the guess is usually "price" — which is what departing customers say when they do not want a longer conversation, in exactly the way buyers say it during a sale.

The uncomfortable summary

Optimization effort in most companies is allocated in almost exactly inverse proportion to what the arithmetic recommends. Acquisition gets the budget, the headcount and the quarterly initiative. Realisation gets a discount-approval policy nobody enforces. Retention gets a customer success team measured on renewal and given no authority over the product decisions that determine it.

None of that is irrational at the level of any individual decision. It is what happens when spending is organised around departments and revenue leaks between them — which is, precisely, the gap a revenue operations function exists to close, and the reason the four numbers above so rarely sit on one person's desk.

The fix is not a better acquisition programme. It is a single afternoon with the four numbers, and the willingness to act on the answer even when the answer is that the cheapest available money has been sitting in your discount policy the whole time.

Common questions about revenue optimization

What is revenue optimization? Increasing revenue by finding and reducing the places it leaks, across the whole chain — acquisition, conversion, realised price, and retention and expansion — rather than by increasing volume at the top. In practice the term is usually used to mean acquisition alone, which is the narrowest and most expensive reading of it.

Where do most companies lose the most revenue? In realised price and in retention, because those are the two leaks that are least instrumented and have the least clear ownership. That is not a universal claim about every business — it is a prediction about where you will find the surprise when you size all four in your own numbers, and it is usually right for companies past their first few years.

Is a ten percent gain really worth the same at every leak? For the three that produce new business, yes, and it falls out of the structure rather than the example: new business is a product of opportunity count, win rate and realised price, so a ten percent change in any factor changes the product by ten percent. Retention breaks the symmetry because it applies to the installed base instead, which is usually bigger, and because it recurs.

Why is net revenue retention worth more than the others? Two reasons. It applies to the base rather than to new business, and for most established companies the base is the larger number. And it persists — an improved retention rate applies to next year's base as well, which includes this year's new business, so the gain compounds rather than needing to be re-bought.

Should an early-stage company follow this ordering? No, and the article says so plainly. With a small installed base there is little for a retention point to apply to, and acquisition really is the right obsession. The ordering flips somewhere around the point where the base exceeds annual new business, and the error most companies make is not noticing for several years after that.

How do I find discount leakage? Report realised price against list by rep and by segment, then look at the shape rather than the average. Uniform discounting across everyone means your price is wrong and no amount of negotiation coaching will fix a positioning problem. Discounting concentrated in a few reps means the price is fine and you have a skills gap.

Is not gross retention just a product problem rather than a sales one? Largely, which is exactly why it gets skipped in sales-led revenue programmes — the work sits outside the department running the initiative. That is an organisational reason, not an arithmetic one, and it is expensive: the biggest number in the model is the one nobody in the room owns.

What if I can only do one thing this quarter? Enforce discount approval and report realised price by rep. It costs nothing, it shows up within a quarter, and it produces the instrumentation you need to argue about everything else. Start the retention work the same quarter anyway, because it will not pay for a year and the year starts whenever you begin.

A note on sources

This article publishes no benchmark for any of the four leaks — no typical win rate, no average discount depth, no standard net revenue retention, no churn figure — and the omission is deliberate on a topic where all four are quoted constantly.

Retention benchmarks are the clearest case. Gross and net retention can be computed on revenue or on logos, annually or monthly, with or without downgrades counted as churn, including or excluding customers still inside an initial term. Those choices move the headline figure by more than the difference between a healthy company and a failing one, and published benchmarks rarely state which convention they used. Discount-leakage research has a subtler version of the same problem: it depends entirely on what "list price" means at a company that negotiates every deal, which is most of them.

The model in this article uses invented numbers, labelled as such, and its findings are properties of the arithmetic rather than of the values. That the three new-business leaks yield identical gains is true at any inputs, because it follows from multiplication. That retention outperforms them depends only on the installed base exceeding new business, which you can check in about a minute.

The framing of revenue as a chain with losses at each joint is standard in pricing and revenue-management literature; net revenue retention as the compounding term is standard in recurring-revenue analysis. The four-leak split, the equal-gains reframe and the sequencing rule are ours — a practitioner's way of deciding where to spend next quarter, which is the part most writing on this subject replaces with a list of tactics.