What data does Revenue Intelligence use?

Revenue Intelligence uses the data you already hold about your revenue: deals and customer details from the CRM, pricing terms from contracts, delivered work from orders, hours or usage, invoice lines from billing and receipts from the accounting system. It is not about having as much data as possible, but about the fields that let you establish what should have been invoiced and whether it was. At a minimum you need CRM deals and invoice lines, with something that lets you link them.

Below, source by source, are the fields that matter, why they matter, and what is usually missing in practice. For the overall picture of the discipline, see the complete guide to Revenue Intelligence.

The three kinds of data

All the data Revenue Intelligence uses falls into three groups:

1. Agreement. What was sold and on what terms? Deals, quotes, contracts, pricing agreements.
2. Delivery. What was actually supplied? Orders, hours, licences, usage, additional work.
3. Money. What was invoiced and received? Invoice lines, credit notes, payments.

Revenue leakage is the difference between these three. An agreement without an invoice. Delivery beyond the agreement without an extra invoice. An invoice without a payment. To find those differences, you need something from each group.

What data do you need for Revenue Intelligence?

CRM: the agreement as sales knows it

The CRM is usually the first place an agreement is recorded. The fields that matter:

• Customer, with a unique key. Preferably a company registration number or a customer account number that also appears in the accounting system.
• Deal status and close date. When was the deal won, and so from when should invoicing have started?
• Deal value, ideally split into one-off and recurring.
• Product lines, with quantity, price and discount per line.
• Term and start date for recurring revenue.
• Owner. Who is responsible if something is wrong?

What is usually missing: product lines. Many salespeople fill in only a total amount. You can then check whether there is an invoice, but not whether the right products were invoiced at the right price. The split between one-off and recurring is often missing too.

It is important to realise that CRM data is not financial data. A deal value is an expectation, not revenue. Why that distinction matters is explained in why CRM data is not the same as financial data.

Contracts: the agreement as it legally stands

The contract is the source for what you are entitled to invoice. The fields that matter:

• Price per unit and the unit itself: per month, per user, per site, per hour.
• Term, start date, end date, renewal arrangement and notice period.
• Indexation clause: which index or which percentage, on which date, with what minimum or maximum.
• Volume tiers: at which quantities does the price change?
• Discounts with an end date: an introductory discount that lapses after twelve months.
• Rates for additional work, urgent work or extra services.
• Minimum volume or fixed fees.

What is usually missing: a system that holds these fields. In many companies the contract exists only as a PDF in a folder or in an e-signature tool. The indexation clause is on page 7 and no system knows about it. That is exactly why leaks between contract and invoice run for so long, as set out in revenue leakage between contract and invoice.

With AI you can now extract these fields from PDFs. Have a person check the extracted values before they are used as the benchmark, especially where clauses are non-standard.

Orders, hours and usage: what was actually delivered

This source varies most by business model:

The fields that always matter: customer, period, quantity, and whether it falls inside or outside the agreement. That last one is crucial for additional work. Hours that were worked as additional work but not flagged as such disappear into the total.

What is usually missing: the link to the customer or contract. Hours are booked to a project number that exists nowhere else, or usage is recorded against a technical account ID that is not linked to a customer account.

Billing: what was charged

From billing, whether that is Exact, AFAS, Xero, NetSuite, SAP or a separate billing tool, you need:

• Invoice lines, not just invoice totals. With description, item or product, quantity, unit price and discount.
• Invoice date and the period the invoice relates to.
• Customer account, with a key that also appears in the CRM.
• References to order, contract or project, where they are passed along.
• Credit notes, linked to the original invoice.

What is usually missing: references. An invoice that carries the contract number or order number can be connected in one step. An invoice with only a free-text description calls for guesswork.

Payments: what came in

The final step. From the accounting system or the bank feed:

• Amounts received, linked to the invoice.
• Open items and their age.
• Write-offs and bad debts.
• Payment behaviour per customer over time.

This is where a different kind of leak sits: not too little invoiced, but invoiced and not received. Small residual amounts that are never followed up, payments that stay just below the invoice amount, write-offs without a clear reason.

Additional sources

Depending on your business, these sources can add value too:

• Support tool: tickets per customer, especially if support outside a contract is invoiced or if rising ticket volumes are an early warning of cancellation.
• Advertising platforms and analytics: to trace spend through to invoiced revenue per channel.
• Price lists: as a reference for what a customer should have paid in the absence of a specific agreement.

What you need at a minimum

To start, you need two things:

1. Won deals from the CRM, with customer, date and amount.
2. Invoice lines from billing, with customer account, date and amount.

And one thing to connect them: a shared customer key, or enough overlap in name and amount to match them reliably.

With that you can already run the check that often yields the most: which won deals have no invoice, or too low an invoice? Contracts, hours and payments come after. Which systems to connect for this is covered in which systems to connect for Revenue Intelligence.

Data quality: good enough is good enough

A common reason not to start is that the data is not in order. Duplicate customers in the CRM, deals without product lines, invoices without a reference. That is the case almost everywhere.

The good news: Revenue Intelligence is one of the best ways to improve data quality. It shows where matching fails, and every failed match is a concrete data gap to close. Waiting until the data is perfect means never starting. What poor data itself costs is covered in revenue leakage from data problems.

Worked example: suppose you have 400 active customers. For 340 of them the same customer account number appears in the CRM and the accounting system. For 60 it does not. You can then compare 85 percent straight away and you have a list of records to clean up for the remaining 15 percent. That is not a reason to wait. It is a to-do list.

Checklist: is your data ready?

Go through these points for each source. Every no is something to improve, not a blocker.

• Do the CRM and the accounting system share a customer key, such as a customer account number or company registration number?
• Do won deals have product lines with quantity and price, or only a total amount?
• Do deals distinguish between one-off and recurring value?
• Are indexation, volume tiers and discounts with an end date recorded anywhere other than in a PDF?
• Do invoices carry an order, project or contract number?
• Is additional work recorded as a separate type in hours or orders?
• Are credit notes linked to the invoice they correct?

How data is protected when you connect it is described by AutoMaat on the security page.

Frequently asked questions

Which data matters most for Revenue Intelligence?

The combination of won deals from the CRM and invoice lines from billing. Without those two there is nothing to compare. Contracts are the most important next source, because that is where the pricing terms are.

Do I need personal data?

Hardly any. Revenue Intelligence works with companies, amounts, dates and products. Names of contact persons are rarely needed. The owner of a deal or customer is useful, though, so a finding can be routed to the right employee.

How much history do I need?

At least twelve months, so you see a full cycle, including an indexation date. More history helps when looking for patterns, such as customers who gradually buy less.

What if my CRM is messy?

Start anyway. The connection between CRM and billing shows exactly which records are messy. That makes cleaning up more targeted than a general clean-up exercise.