Hail Mary, but Hubspot Insights enriched company data= hot garbage, want to undo it

[SaleProcessNerd]

Hey all,

So… Hubspot Insights (Get started with data enrichment) had the, conceptually, great idea of auto-populating empty fields on a Company record:

• Annual revenue
• City
• Country
• Description
• Industry
• Is public
• Name
• Number of employees
• Phone number
• Postal code
• State/Region
• Street address
• Street address 2
• Time zone
• Total money raised
• Website URL
• Year founded
• Facebook company page
• LinkedIn bio
• LinkedIn company page
• Twitter handle

With whatever data they had available in the ‘Insights’ database. Unfortunately, the ‘fill-in-the-blanks’ data in the Hubspot Insights database is flaming hot garbage, and now I have hundreds of company records with gems such as:
Located in:

city: “San Francisco”

state/region: [u]

“Ariana”

[/u]

country/region: [u]

“Tunisia”

[/u]

and

city: “Vancouver”

state/region: “B.C.”

country/region:

“Montenegro”

(If you’re less N.A. centric, San Fran California and Vancouver BC are well known, large cities that DO NOT exist in Tunisia and Montenegro (they are in U.S.A. and Canada respectively)
Anywhoo…
On individual Company records, it does show whether that data was popualted by Hubspot Insights:

so it stands to reason that Hubspot, on some level, had a way to ‘flag’ all fields populated by steaming, useless garbage Hubspot Insights.
Any of you fellow nerds out there aware of any way to pull a list or report or identify which of my company records, or individual fields, were graced with the inverse midas touch of the Insights database?
Will send a free 12" pizza to anyone that can actually solve this. Not joking (assuming I can call/order using something like ‘Uber Eats’/‘Skipthedishes’ etc…
plshelp

[KyleJepson]

A partial answer:

First, there’s a setting where you can turn off HubSpot Insights. That won’t remove the existing values that have already populated, but it’ll prevent any future records from having insights added to them. More on that here.

Next, when you view a property’s history, you can see who it was updated by, and that includes HubSpot Insights. So that information is on the record. Which brings me to some good news and some bad news:

Good news: You can pull this information through the API.

Bad news: The most recent version of our API (v3) doesn’t support this yet, so you need to use the older API (v2)--which means you can’t do this through a custom-coded workflow action, because those use v3 (well, they use HubSpot JS Node, which uses v3).

So that’s why this is a partial solution--because I’m quickly going to get out of my depth here. But I dug through the legacy API docs, and I found this for you:

https://api.hubapi.com/companies/v2/companies/paged?hapikey=HAPIKEYGOESHERE&properties=annualrevenue&properties=city&properties=country&properties=description&properties=industry&properties=is_public&properties=name&properties=numberofemployees&properties=phone&properties=zip&properties=state&properties=address&properties=address2&properties=timezone&properties=total_money_raised&properties=website&properties=founded_year&properties=facebook_company_page&properties=linkedinbio&properties=linkedin_company_page&properties=twitterhandle

That’s the URL you need to hit with an http GET. It will pull the 21 properties updated by insights with their sources for all of your companies. Replace HAPIKEYGOESHERE with your actual API key, and that will pull the information you need.

When you dig through the response from that call, you’ll see that each property has a Source and a Source ID, like this:

Source ID, I don’t really understand. Not reliable. Ignore. But Source--that’s what you’re looking for. If it says BIDEN, then the data came from HubSpot Insights. (I don’t know why, so don’t ask.)

So if this worked with custom-coded workflow actions, you could probably write some code that would pull in a company, look at those 21 properties, and for the ones with a Source of BIDEN, delete the data. But like I said, this doesn’t currently work with custom-coded workflow actions, so a different solution is needed.

And that’s where my helpfulness hits a limit

[SaleProcessNerd]

Thanks Kyle.
I think I might be OK accessing older versions of the Hubspot API, so this is super helpful. If I can get this to work, message me your preferred delivery address, pizza, and date and i’ll send over a 'za for ya.

mmmm slaveBots