Self-serve signups and a sales team, arguing over the same customer

A software company runs a free trial that anybody can start and a sales team that closes annual contracts. Marketing reports signups by channel. Product reports activation and conversion. Sales reports closed-won by opportunity source, populated by a salesperson from memory. Finance reports recurring revenue. Four systems, four numbers, and no path between the advertisement somebody clicked and the contract somebody signed.

Product-led and sales-led businesses each have a measurement story that works. A business running both has two stories that contradict each other and no arbiter between them.

This is the normal condition for mid-market software companies, and the arguments it produces — about who owns the pipeline, whether marketing sourced a deal, whether product-qualified leads are real — are usually definitional rather than factual.

Attribute the account, not the signup

The unit of attribution in a hybrid SaaS business is the account, not the individual signup. One account may produce several signups, from several people, months apart, before any money changes hands.

A typical enterprise journey involves an engineer starting a trial out of curiosity, a colleague signing up separately having heard about it internally, and a procurement process beginning much later with a third person who never used the product. Attributing at the signup level treats those as three unrelated events and connects none of them to the eventual contract.

Account matching is what joins them, usually on the email domain, sometimes on a company identifier where one is captured. Domain matching is unusually effective in software because business users overwhelmingly sign up with work addresses, and it is the single highest-value change most SaaS attribution setups can make.

The exception to watch is generic domains. A signup from a personal address cannot be domain-matched, and in categories selling to small businesses that may be a large share of your base. Those accounts need person-level matching, and mixing the two populations without noticing produces misleading conclusions about channel quality.

Recurring revenue, at a fixed horizon

Attribute annual recurring revenue at a fixed point after acquisition rather than the first payment. First-invoice attribution undervalues every channel that brings accounts which expand and overvalues every channel that brings accounts which churn.

On first invoice, paid search is the best channel and content is third. At twenty-four months the ranking has completely inverted. A business optimising on first invoice value would have spent two years moving budget away from its two most valuable sources.

This is not an unusual pattern in software; it is close to the default one. Channels that capture immediate intent tend to bring buyers with immediate needs and shallow commitment, while channels that build understanding tend to bring accounts that fit better and stay longer.

Expansion and contraction, credited properly

Expansion revenue belongs to the channel that acquired the account, reported separately from new business. Ignoring expansion makes acquisition look less valuable than it is and makes it impossible to compare channels on their real economics.

The reporting convention that works is straightforward: attribute new business to the acquisition source at the point of first payment, then report each cohort's net revenue at fixed horizons, showing expansion and contraction as their own lines. That preserves the acquisition credit without pretending marketing caused the expansion.

Whether marketing deserves credit for expansion is a genuine question with no universal answer, and it is a distraction. What matters for allocation is that a channel produces accounts that grow, whoever gets the credit internally, and reporting the fact does not require settling the ownership argument.

Product-qualified leads and the definition problem

Most disputes about product-qualified leads are disputes about definition, not about data. Write the definition down, validate it against closed revenue, and re-validate it whenever the product changes.

The usual failure is that a product-qualified lead is defined by a usage threshold chosen intuitively, never checked against outcomes, and then relied on for years while the product changes underneath it. A threshold that predicted conversion when it was set can become meaningless after an onboarding redesign, and nothing in the reporting will announce that.

Validation is not difficult. Take accounts that crossed the threshold and accounts that did not, and compare closed revenue over the following year. If the difference has narrowed, the definition has drifted and needs resetting. Doing this twice a year is enough.

The opportunity source field, and why it is unreliable

In most SaaS CRMs the opportunity source is populated by a salesperson at the point of creation, from memory, under time pressure. It is the weakest data in the reporting chain and the one most decisions rest on.

This is not a criticism of sales teams. Asked to record where an opportunity came from, a person will name the most recent or most memorable interaction, which is frequently a demo request rather than the eighteen-month sequence of content and events that preceded it. The field is doing what a human can do.

The remedy is to keep it and stop relying on it. Retain the self-reported source as its own field, match independently against marketing records, and compare. Disagreement between the two is informative — it usually indicates where the recorded journey is incomplete — and the comparison is far more useful than either alone.

It is worth being specific about how the field goes wrong, because the pattern is consistent and fixable at the margin. Opportunities created from an inbound demo request are recorded as inbound; opportunities created after a salesperson follows up on a trial are recorded as outbound or sales-generated, even when the trial itself came from a marketing channel. The field is capturing who created the opportunity record rather than where the account came from, and those two questions have been quietly collapsed into one dropdown in most implementations.

CloseRev matches on account as well as person, and shows the source records behind every match, so a self-reported opportunity source can be checked against what the marketing systems actually recorded rather than accepted as given.

Free users who never pay, and what they are worth

In product-led businesses a large share of signups never pay anything, and treating them as failed acquisitions misses their role in referral, community and eventual conversion within the same account.

The account-level view handles this correctly and the signup-level view does not. Three free users at one company followed by a paid contract from that company is a successful acquisition; counted individually it is a seventy-five percent failure rate and one conversion.

It is worth measuring the referral effect too, where the product supports it. A free user who never pays but introduces the product into two other organisations is producing revenue that no direct attribution will ever assign to them, and businesses that cannot see this systematically underinvest in the free tier.

Trials, and the attribution window nobody sets

The gap between a signup and a first payment can be days in a self-serve motion and eighteen months in an enterprise one. Running both through a single attribution window will systematically favour whichever motion is faster.

This is a quiet and consequential error. A ninety-day attribution window applied across the whole business will capture nearly all self-serve conversions and almost no enterprise ones, so the channels feeding the enterprise motion will appear to produce signups that never convert. Budget then moves toward the faster motion, which shrinks the pipeline for the slower and higher-value one.

The fix is to run separate windows by motion, determined by the actual observed distribution rather than by a round number. Most businesses have never calculated the time from first touch to first payment separately for each motion, and the two figures are usually far further apart than anybody assumes.

It also affects how quickly a marketing change can be judged. A shift in content strategy aimed at enterprise accounts cannot show revenue evidence for a year or more, and evaluating it on a quarterly cadence against self-serve conversion data will reliably kill it before it works.

Attribution for a product that spreads by itself

Where a product spreads through invitation, sharing or collaboration, a meaningful share of new accounts arrive with no marketing touch at all. That is a genuine acquisition channel and it needs reporting as one rather than sitting in an unattributed bucket.

Viral or collaborative acquisition is measurable from inside the product, because the invitation carries a source. What usually prevents it appearing in marketing reporting is that the data lives in the product analytics system and never reaches the revenue view, so it is invisible in exactly the report where budget is decided.

Bringing it in changes the picture in two directions. It reduces the apparent unattributed share, which improves the credibility of the whole report, and it makes visible that some paid channels are being credited with revenue that arrived through a colleague's invitation rather than through the advertisement.

It also creates a useful strategic measure: the ratio of invitation-driven accounts to marketing-driven ones, tracked over time. A rising ratio is one of the strongest signals available that the product is working, and it is the kind of number a board will remember.

The pricing change that invalidates your history

Software businesses change pricing and packaging regularly, and each change breaks comparability of revenue-per-account across cohorts. Record pricing changes alongside the attribution history or the trend will be misread.

A cohort acquired before a price increase and one acquired after are not comparable on revenue per account, and the difference will present as a channel performance change if nobody is tracking the pricing timeline. Businesses that repackage every year effectively reset their cohort comparability every year without noticing.

The remedy is a simple annotated timeline maintained alongside the reporting: what changed, when, and which cohorts it affects. It takes minutes to maintain and prevents the most common misinterpretation in software attribution, which is attributing a pricing effect to a marketing channel.

Where a change is large, consider reporting cohorts on units rather than revenue as well — accounts acquired, seats acquired — since those remain comparable across pricing regimes and often tell the clearer story about channel performance.

What to report to a board

Report new annual recurring revenue by acquisition channel at a fixed horizon, net revenue retention by acquisition channel, and the share of revenue with no traceable source. Three figures cover almost every question a software board asks.

Net revenue retention by acquisition channel is the item most boards have never seen and find most useful, because it separates channels that acquire well from channels that acquire durably. Those are frequently not the same channels, and the difference compounds over the periods a board is thinking about.

The unattributed share matters here for the same reason as everywhere else, and in software it is often larger than expected because so much evaluation happens in communities, peer recommendations, podcasts and content nobody clicked. Stating it protects every other number on the page, and its direction over time is a genuine signal in its own right: a shrinking unattributed share usually means your instrumentation improved, while a growing one in a growing business often means word of mouth is doing more of the work than your reporting can see.

The organisational fix

The measurement problem in hybrid SaaS businesses is usually organisational: marketing owns acquisition data, product owns behavioural data, sales owns the opportunity record and finance owns revenue, and no single view joins them. Somebody has to be accountable for the joined view.

Naming that owner is more consequential than any tool decision. Without one, each function optimises its own metric — signups, activation, pipeline, recognised revenue — and all four can improve while the business does not, because nothing in the arrangement requires them to be consistent with each other.

The owner does not need to be senior and does need authority to define terms. Most of the value is unlocked by a single agreed definition of an account, a source and a revenue figure, applied consistently across four systems that currently each hold their own.

There is a version of this that fails, and it is worth naming so it can be avoided. Appointing an owner for the joined view without giving them access to all four systems produces somebody accountable for a number they cannot compute, which is worse than having nobody, because it creates the appearance of ownership while the underlying disconnection continues. Access to the revenue data in particular is the one that gets withheld, usually for good reasons that nobody revisits.

The practical test of whether the fix has worked is simple: can one person, in one afternoon, produce a table of annual recurring revenue by acquisition channel that finance recognises? If the answer requires a project, the ownership is nominal. If it requires four meetings, the definitions have not actually been agreed, whatever the documentation says.

Four teams, four dashboards, four numbers that all went up. Somebody has to own the one number that says whether the business did.

Questions people actually ask

How do you attribute revenue in a SaaS business with both self-serve and sales motions?

Match at the account level rather than the signup level, and attribute recurring revenue rather than the initial conversion. A user who signs up on a free plan and a procurement officer who signs an annual contract eighteen months later are different records belonging to the same customer.

Should SaaS attribution use new ARR or first invoice value?

Annual recurring revenue at a fixed point after acquisition, typically twelve months, because it captures expansion and contraction. First invoice value systematically undervalues channels that bring accounts which grow.

How do you credit marketing for expansion revenue?

Attribute expansion to the original acquisition source and report it separately from new business. A channel producing accounts that double in year two is more valuable than one producing accounts that churn, and only cohort reporting shows the difference.

Who gets credit when a self-serve user becomes an enterprise deal?

Both, reported separately. The acquisition source brought the account; the sales motion converted it. Forcing a single owner produces an argument that has no correct answer and consumes more time than the decision is worth.

Does product analytics replace marketing attribution in SaaS?

No. Product analytics explains behaviour inside the product and generally has no view of what was spent to acquire the account or what it eventually paid. The two answer different questions and are usually owned by different teams, which is why they rarely get joined.

How do you handle churn in attribution reporting?

Net it against the acquiring channel at a fixed horizon. A source producing customers who churn within six months is not performing as its first-month revenue suggests, and gross acquisition reporting will rank it far too highly.

Found this useful?

• Share on X
• Share on LinkedIn
• Share on Reddit
• Share on Email