csv file import mismatch

[JSPrenovost]

We recently imported thousands of companies into HS using CSV files. Mapping was done correctly and source files were automatically generated from a tool that scraped LinkedIn. 98% of companies were imported correctly (associating the right domain and letting HS fill in the available info).

Oddly some companies were assigned another domain and therefore all information on them is flawed. Searching for the company whose domain was incorrectly used, we found that the domain used for that company was also wrong, and so on until a domain used was not part of our ABM list.

I looked into the error files and there was no trace of the companies in the import error files. (there were not involved of a copy or update error, etc) 
Information in the source file was correct and structured correctly.

Any clues as to why the import was almost correct at 100% but not? We will have to find out the culprits manually and change the domain before scraping contacts and start sequencing and prospecting.

Any help would be great.

[franksteiner79]

Hi @JSPrenovost 

Without knowing what information you imported it's a bit difficult to assist.

Just to rule out the obvious, are you 100% sure that those wrong companies were part of the import? You can check that by looking at the companies record source.

If you don't import a domain but a website URL (these are two different things in HubSpot), HubSpot will try and work out the company domain from the website URL, that can sometimes lead to a mismatch.

Frank

[JSPrenovost]

100% sure that the companies were created with the import as the record source.

I only used 2 columns in the original csv file and mapped them with HS (Company name and domain) for the import. As an example, in the CSV, a company called October's very own with it's domain  https://octobersveryown.com 

The import domain was infoq.com. That domain was in the file on anotehr line with InfoQ company. Then InfoQ company import had the same issue and so on with 6 other companies.

As mentioned, the problem here is that 98%+ of all imported companies were ok and even though data is consistent within the original CSv, it imported a few random ones wrong.