The most expensive data problem has no owner
Two AEs call the same CEO in the same week. Both have an account in the CRM, one “Müller GmbH”, one “Mueller GmbH & Co. KG”, both with an open opportunity. On the second call the CEO dryly asks whether your colleagues actually talk to each other. From that moment on, the deal is an uphill climb.
That’s the visible version of the problem. The invisible one is more expensive.
Your attribution falls apart because the marketing contact hangs on duplicate A and the won deal on duplicate B, which means the campaign that brought in the customer shows up in the report as a cost center with no results. Your existing customer gets the new-prospect email saying “Get to know us” because his second record never got customer status, and replies with a screenshot of his last three invoices. Your scoring, your routing, your segmentation: all of it computes on a data foundation where the same company exists two, three, sometimes five times. Every automation becomes a coin flip.
Why is this, of all things, the most expensive data problem? Not because it’s the biggest. But because it has no owner. Sales says data hygiene is marketing’s job. Marketing says the accounts belong to sales. The admin, if there is one, has 40 other tickets. So nothing happens, for years, and everyone gets used to trying three search variations before every call.
At some point comes the spring cleaning. An intern, an Excel list, two weeks of merging. Afterwards it’s better. Six months later it’s all back, because the sources keep running: the web form with no matching, the import from the trade show list, the enrichment tool cheerfully creating new records, manual entry with no required checks. Duplicates aren’t a backlog you clear once. Duplicates are an inflow.
That’s why dedupe isn’t a cleanup project but ongoing governance. Matching rules that fit your data, German legal forms included, no standard ruleset cleanly matches GmbH against GmbH & Co. KG. Matching at every entry point, not after the fact. Clear merge logic, which record survives and which fields win. And above all: a name. One person who is accountable, with a report showing the duplicate rate per month.
Go search for your biggest customer in the CRM. Once spelled out, once abbreviated, once with the umlaut. How many hits?
