Revenue leakage from data problems

Revenue leakage from data problems arises in two ways. Directly: a wrong price in the master data, a duplicate customer where one record is missing the contract, or an empty field that blocks an invoice line costs money outright. Indirectly: bad data makes every other leak invisible, because you cannot reliably put your CRM, billing and contracts side by side. The second is usually the bigger problem. If your data quality is not in order, you cannot even find the other leaks.

Why is data a leak category of its own?

With most forms of revenue leakage the cause is a process: an indexation nobody applies, a job sheet left lying around, a contract that expires. With data problems the cause is the information itself. The process can be perfect, but if the input is wrong, the output is wrong too.

Data problems are also an amplifier. An invoicing error you could have found in a minute with good data goes unnoticed for years with bad data. That is why data quality is not an IT topic but a financial one. It is the precondition for every check.

Which data problems cost the most?

1. Duplicate customers. The same customer appears two or three times in the CRM or the billing system, under slightly different names: "Baker Ltd", "Baker Limited" and "Baker Building Services". The contract hangs off one record, the invoices off another, the usage off the third. Every comparison at customer level fails. Discounts meant for one record are sometimes applied to all of them. Volume discounts are calculated on part of the revenue instead of the whole, or counted twice.

2. Empty or unreliable fields. The "contract value" field in the CRM is empty on a third of the deals. The "end date" field is filled in on half of them with the date someone created the field. "Deal owner" points to an employee who left two years ago. Every report that relies on those fields is unreliable, and every automated check that uses them raises false alarms or no alarm at all.

3. Wrong master data. The price of an item in the ERP was never updated after the last price increase. A customer is on the wrong price list. A VAT code is set up incorrectly. Because master data is used in every transaction, one error multiplies across hundreds of invoices.

4. No shared key between systems. The CRM knows customer 10482, the accounting package knows debtor 3317, the support system knows an email domain. There is no field connecting the three. Every comparison then starts with manual matching, and what is matched by hand is rarely kept up to date.

5. Conflicting definitions. Sales reports revenue as the value of won deals. Finance reports invoiced revenue. Management looks at a dashboard that uses a third definition. In the board meeting three people quote three figures, and nobody knows which difference is a leak and which is a definition. The question of which data source is leading for revenue therefore has to be answered before you can compare anything.

How do data problems hide other leaks?

Suppose you want to know whether every won deal has been invoiced. You put the won deals from the CRM next to the invoices from the accounting package. Because of duplicate customers and missing keys, 300 deals are left without a match. Perhaps 20 of them genuinely have not been invoiced. The other 280 were invoiced under a different customer name, a different deal number or combined on a single invoice.

Nobody is going to work through 300 lines by hand. So the check is not done, or done once and never again. The 20 real leaks stay hidden behind 280 false alerts. That is the indirect effect: bad data makes checks so expensive that they do not happen.

This is closely linked to manual administration. Manual work causes data problems, and data problems increase the manual work needed to find them.

Worked example

Worked example: suppose a wholesaler with EUR 8 million in revenue discovers the following during a clean-up:

• 60 duplicate customer records. For 5 of them, someone had given the customer a tiered discount on total purchases on one record, while the other record still carried its own customer discount. So the customer received two discounts where one had been agreed. Average difference per customer: EUR 1,500 a year, EUR 7,500 a year in total. For 8 others, the split meant the customer fell into a lower tier and paid too much. That is not a leak, but it is a complaint waiting to happen.
• 45 items with an outdated price in the ERP, on average 6 percent too low, with combined annual revenue of EUR 400,000: EUR 24,000 a year.
• 3 customers on an outdated price list with combined revenue of EUR 250,000, on average 4 percent too low: EUR 10,000 a year.

Together more than EUR 41,000 a year in direct leakage. That excludes what the clean-up makes possible: a comparison between CRM and billing that now actually works, and the leaks it brings to light.

Where do you start?

Data quality can become an endless project. So do not start with "get all the data in order", but with the fields you need to check money.

1. Identify the five to ten fields every revenue check needs. Usually: customer ID, contract value, start date, end date, price, product or package, owner.
2. Measure how often each field is empty or illogical. An end date before the start date. A contract value of zero on a won deal. An owner who no longer works for the company.
3. Create one shared customer key. Choose the leading system, usually the accounting package or the CRM, and put that ID into the other systems. This is the single most important piece of data work you can do.
4. Merge duplicates. Start with the customers with the highest revenue. When merging, check which contracts, discounts and arrangements were attached to which record.
5. Check the master data used in every transaction. Price lists, item prices, discount structures, VAT codes. Compare them with the most recently approved prices.
6. Write down definitions. What is revenue, what is an active customer, what is a won deal? Put it in writing, and use the same definition in every report.

How do you keep it clean?

A clean-up is polluted again within six months if nothing changes in how data comes in.

• Mandatory fields at the source. A deal cannot be marked as won without a contract value, start date and customer ID. Most CRMs, such as HubSpot, Salesforce and Pipedrive, support this with validation rules.
• A duplicate check on creation. Warn when a new customer resembles an existing one, by name, company registration number or email domain.
• Periodic quality measurement. Measure the same short list every month: percentage of empty core fields, number of possible duplicates, number of records without a shared key. A rising number is a signal.
• An owner for each system. Someone responsible for data quality in the CRM, and someone for the ERP. It does not have to be a full-time role, but it must be a named task.

The use cases page lists data quality as one of the eight areas where revenue leaks, with what that pattern looks like.

Frequently asked questions

Does bad data really cost money, or is it just inconvenient?

Both. Wrong prices and discounts in the master data cost money directly. But the largest effect is indirect: bad data makes checks so cumbersome that they are not done, so other leaks persist.

Do I have to clean up all my data before I can look for leaks?

No. Start with the fields you need for one check, for example won deals against invoices. Clean up whatever blocks that check. That way every piece of clean-up work pays off straight away.

What is the most important field to get right?

A shared customer key that is the same in every system. Without it, every comparison between systems is manual work. With it, most other checks become simple.

How do I stop the data getting polluted again after a clean-up?

By enforcing quality at the source: mandatory fields, a duplicate check on creation and a named owner for each system. Cleaning up without those measures is mopping the floor with the tap still running.