Pick one canonical form for a company name, write it down, and enforce it at the point of entry. Standardising what is already in the CRM without that rule is a job you will do twice, because tomorrow's records arrive through the same imports, the same web forms and the same reps that created the mess in the first place.
Company names deserve more care than most fields because so much hangs off them. Reporting by account, deduplication, enrichment, territory assignment and any integration that matches on name all quietly assume "Acme Ltd" and "Acme Limited" are the same company. To the CRM, they are two.
Agree the canonical form before touching the data
The canonical form is the one version of a name the CRM treats as correct. Most of the disagreement lands on four things, and all four are worth settling explicitly:
- Legal suffix: keep "Ltd", drop it, or spell out "Limited". Dropping it reads best in reports, but it removes the only thing separating two unrelated firms with the same short name.
- Punctuation: "Smith & Sons" or "Smith and Sons", whether a comma precedes the suffix, whether initials carry full stops.
- Case: sentence case applied to everything, or the name kept as the company styles it. "BAE Systems" and "innocent drinks" both break a blanket rule.
- Leading words: "The Acme Group" or "Acme Group". Sorting and alphabetical browsing depend on it.
None of these choices is right in the abstract. What matters is that one of them is written down somewhere a new starter can find it, because the cost sits in the inconsistency rather than the convention.
Trading names belong in a second field
The most common mistake is making one field carry two facts. A business invoices as "Northern Components Holdings Ltd" and everyone in its market calls it "Nortec". Put either name in the company name field alone and half your team cannot find the record.
Keep the legal name as the canonical company name and give the trading name its own field. Search then works on both, reporting stays consistent, and nobody has to decide which name is more useful this week. The same applies after a rebrand: a field holding the former name is what stops someone creating a fresh record for a company you have worked with for years.
Groups and subsidiaries need a relationship, not a naming convention
Teams often encode structure into the name itself, as "Acme Group (Scotland)" or "Acme - North". It works until someone needs to report on the group as a whole, at which point the structure exists only as a string nobody can filter on reliably.
HubSpot and Pipedrive both support linking company records to one another: parent and child companies in HubSpot, related organisations in Pipedrive. Use it. A subsidiary should be named the way it trades and linked to its parent, so the group view is a query rather than a text pattern someone has to remember.
Automate the tidy-up, but do not let it merge unattended
Most of the backlog can be normalised programmatically: trimming stray whitespace, unifying suffixes, fixing shouted names, collapsing double spaces. That work is mechanical, repeatable and safe enough to script.
What is not safe is letting a normalisation pass merge records on the strength of a name match. Standardising "Acme Ltd" and "Acme Limited" makes them look identical, which is the point, and also the risk: two genuinely unrelated firms that share a short name now collide. Domain is the better tiebreak, because companies sharing a name rarely share a website.
The practical sequence:
- Normalise the formatting first, as a bulk pass, against rules already agreed.
- Review the matches that normalisation creates, checking domain and address rather than name.
- Only then merge, with deliberate rules about which record survives.
The review step is not optional. Automated matching is right most of the time, and the minority it gets wrong costs more to unpick than the whole exercise saved.
Hold the standard where records are created
The rules only hold if the CRM makes them easy to follow. In practice that means a stated format on the company name field, a duplicate check that fires when a record is created rather than at the next audit, and imports validated against the same rules before they load rather than after.
Without that, the standard decays at the rate new records arrive, which is exactly why data cleaning projects stop sticking.
The pattern, in short
Standardising company names is one decision and one habit. Agree the canonical form, covering suffix, punctuation, case and leading words, then enforce it where records are created. Trading names get their own field, group structure gets a relationship rather than a naming convention, and automated normalisation gets a review step before anything merges.
The limit worth being honest about: no rule survives every real company name. Some records will always need a human to judge them. The aim is an exception list short enough that judging them is a few minutes' work rather than another project.
Company names all over the place in your CRM?
Get in touch. We will agree the rules with you, run the clean-up and put the check at the entry point so it holds. That is data cleaning work, and it is most of what we do.