Home / CRM / RevOps & Attribution
CRM
RevOps & Attribution
Connect marketing, sales and data so every lead is tracked to revenue.
Overview
What we do & why it matters
RevOps is the connective tissue that makes marketing measurable. We unify your data, define the funnel, and build attribution so you can see which content, ranking or AI citation actually drove the deal.
What’s included
- Funnel & data-model design
- Multi-touch attribution
- Lead scoring & SLA design
- Reporting & forecasting dashboards
- Tool consolidation & integration
- Ongoing RevOps support
Part of our CRM engagement · from transparent pricing.
The problem
Three teams, three versions of the number.
RevOps exists because marketing, sales and finance each measure revenue with their own definitions, sources and timing. The work is less about tooling than about agreeing on definitions and then enforcing them in the systems.
Shared definitions
Qualified, opportunity, won, churned — written once, used everywhere, or every report needs a footnote explaining itself.
One attribution model
Pick a model, document its limits, and stop switching. A consistent imperfect model beats three perfect contradictory ones.
Handover as a process
Where marketing ends and sales begins is a policy decision. Undefined, it becomes an argument repeated monthly.
Forecast hygiene
Forecasts built on stages without exit criteria are opinion polls. Fix the stages and the forecast improves without new software.
The outcome is unglamorous and immediately useful: everyone reads the same dashboard and argues about the business instead of about the data.
Scope
What an engagement covers.
Foundations
The definitions everything else depends on
Most revenue reporting arguments are definitional rather than technical. Two teams counting different things will never reconcile their numbers, however good the dashboard is, and no tooling resolves a disagreement about what a qualified lead means.
Stage exit criteria
What must be observably true for a deal to advance. Without this, stage data is opinion and the forecast built on it is a wish.
Qualified, defined jointly
Agreed between marketing and sales, in writing, with examples. The most common unresolved argument in B2B and the most expensive.
One source per field
Which system owns which value. Three systems holding a customer's plan means three answers and no truth.
Account and contact hierarchy
How subsidiaries, divisions and multi-company contacts are handled. Unglamorous, and the cause of most reporting discrepancies at scale.
Consistent time definitions
When a deal is counted, in which period, under which currency conversion. Small decisions that make or break a board number.
Written and accessible
A definitions document people can actually consult. Shared understanding that lives in three people's heads is not shared understanding.
This work is unglamorous and it is the reason reporting projects succeed or fail. We do it first and we put it in writing.
Attribution
A model you can defend to a CFO
Every attribution model is wrong in a specific way. The useful question is not which is correct but which is wrong in ways you can live with and explain — and whether everyone has agreed to be wrong the same way.
First touch
Credits discovery, ignores everything that closed the deal. Useful for judging awareness channels, useless for budget allocation.
Last touch
Credits the final step, usually branded search or a direct visit. Flatters the bottom of the funnel and starves the top.
Multi-touch
Spreads credit across the journey. More honest about reality, and entirely dependent on tracking most touches, which nobody does.
Self-reported source
Asking buyers how they found you. Imprecise, and frequently closer to the truth than the tracking, especially for dark social and word of mouth.
Holdout testing
The only approach that establishes causation rather than correlation. Expensive, disruptive, and the one that settles arguments.
What we recommend
Two or three views reported side by side, with self-reported data alongside, and the disagreements between them treated as information.
We will not supply one confident number. A single attribution figure presented without its assumptions is a political instrument rather than a measurement.
The work
What a RevOps engagement actually contains
Sequenced so the definitional work lands before anything gets built on top of it.
The most valuable output is usually not the dashboard. It is the written definitions document that ends arguments that have been running for years.
Reporting
Dashboards people actually use
Most reporting projects deliver one comprehensive dashboard that serves nobody, because the board, the sales leader and the marketing lead need different things at different frequencies. Building three focused views is less work than building one that compromises between them.
The board view
A small number of numbers with trend and context. Monthly, stable over time, and boring by design — a metric that changes definition is worthless to a board.
The sales leader view
Pipeline coverage, stage velocity, deals at risk and forecast against target. Weekly, and built to prompt action rather than to summarise.
The marketing view
Source performance measured against pipeline rather than leads, with brand and non-brand separated and the attribution assumptions visible.
The individual view
What each person needs to do today. The only dashboard that gets opened daily, and the one most often not built.
Definitions attached
Each metric linked to how it is calculated. A number without its definition restarts the argument it was meant to settle.
Retired when unused
Dashboards nobody opens get deleted. A reporting layer accumulates clutter exactly like a CRM does.
We check what actually gets opened after a quarter and prune accordingly, rather than counting the number of dashboards delivered as the result.
Honest limits
What RevOps cannot settle
Operations improves the quality of information available to decisions. It does not make the decisions, and it cannot measure everything.
Causation without experiments
Attribution describes correlation. Whether a channel caused revenue can only be established with a holdout, and that is a different project.
Untracked influence
Word of mouth, communities, conversations in private channels. Real, material, and largely invisible to any tracking you can buy.
Organisational disagreement
If sales and marketing cannot agree on what qualified means, better tooling produces better-looking disagreement.
A pipeline problem
Perfect visibility into insufficient pipeline is still insufficient pipeline. Reporting does not create demand.
Data nobody maintains
Dashboards inherit the discipline of the people entering data. Where that is absent, the output is confident and wrong.
Forecast certainty
A better forecast is a narrower range, not a number. Anyone promising precision is describing a model that will be wrong with confidence.
More CRM
The rest of the CRM stack.
Questions
RevOps & Attribution, answered.
What attribution model do you use?
We fit the model to your sales cycle — usually multi-touch — so credit reflects reality rather than a convenient default.
Do we need RevOps if we’re small?
Even a lean setup benefits from clean data and clear attribution — it’s cheaper to build right early than to untangle later.
Do we need RevOps if we are small?
You need the definitions, which cost nothing but a difficult afternoon. You probably do not need the tooling, the dashboards or a dedicated function. A ten-person company that has agreed in writing what a qualified lead is and what each deal stage requires has most of the value available to it. The infrastructure becomes worth building when several people are entering data, when marketing and sales are separate functions, or when someone outside the company is asking for numbers you cannot currently reproduce.
What attribution model do you use?
Several, reported alongside each other, with the assumptions of each stated. A single number presented without its caveats is a political instrument rather than a measurement — it gets used to defend budgets rather than to allocate them. We typically show first touch, a multi-touch view and self-reported source together, and we treat the disagreements between them as information. Where a decision genuinely hinges on causation, we will propose a holdout test instead of arguing about models.
Our numbers never reconcile between systems. Can you fix that?
Usually, and the cause is almost never technical. Reconciliation failures come from different definitions — different time treatment, different qualification criteria, different handling of currency or of multi-entity accounts — and from more than one system claiming ownership of the same field. We start by establishing which number is supposed to be authoritative for each question, then make the systems agree. The technical work is the easy half.
How accurate can forecasting get?
More accurate than most companies manage, and never precise. A forecast built on stage exit criteria and historical conversion rates, with accuracy tracked and the model corrected over time, is substantially better than asking representatives how confident they feel. What it produces is a narrower range, not a number. We also report forecast accuracy as a metric in its own right, because a forecast nobody checks afterwards never improves.
Who should own RevOps internally?
Someone with authority across sales, marketing and finance, which usually means it reports to whoever owns revenue rather than sitting inside one function. Owned by marketing, it optimises for marketing's numbers; owned by sales, the same in reverse. In smaller companies it is often a part of the finance or operations lead's role rather than a dedicated hire, and that works provided the authority is real.
Can you work with our existing stack?
Yes, and we would rather fix what you have than propose a replacement. Most RevOps problems are definitional and procedural rather than tooling problems, and buying a new platform to solve them reliably produces the same confusion in a more expensive system. Where a genuine tooling gap exists — usually around data warehousing or attribution infrastructure — we will name it specifically and explain what it is needed for.
How long before this produces something useful?
The definitions document, which is often the most valuable output, comes out of the first few weeks. Data audit and tracking fixes run four to eight weeks depending on how many systems are involved. Usable reporting follows that, and forecast accuracy improves over a couple of quarters as the model gets corrected against actual outcomes. The unglamorous early phase is the part that determines whether the later phases work, so we resist pressure to start with the dashboard.
How does this connect to the paid and SEO programmes?
It is the foundation both depend on. Paid campaigns optimise toward conversion events that come from the CRM, and the quality of those events decides whether the bidding finds buyers or form-fillers. SEO investment is justified by pipeline attribution that comes from the same place. When a company asks us to start with a channel and the revenue data cannot support measuring it, we will argue for doing this first — otherwise you are buying work nobody can evaluate.
Ready when you are
Let's find your next 30% of growth.
A free audit across SEO, GEO, CRM & automation — no strings, no 'contact for pricing'.