-
Notifications
You must be signed in to change notification settings - Fork 377
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Request: Don't use Clearbit on signups from personal domains #8383
Comments
I was about to say yes but actually this would be useful - I think we should start doing this! (Recent relevant context) Basically it could be useful to get alerts purely based on company size, but ignoring ICP scoring stuff. (If the only intention is saving money anyway, we probably don't need to do anything til our Clearbit renewal is up in October?) |
Normally I'd be the one to push this, but that feels really enterprisey. If I Clearbit identifies a user on a personal domain who has signed up but who actually belongs to a big company, there's two likely outcomes in my opinion...
|
I actually think in case number 1 above, if we wait until we see they have strong usage and retention, I wouldn't be grumpy if someone reached out after a few months and said "hey we've noticed you use PostHog a bunch, and our data says you work at X company, if you're struggling with your data stack lmk and we can maybe help." However I have no strong opinion on this! (I do have a strong opinion about somebody renaming their company "X".....) |
For: PostHog/meta#193
This is something I floated in the last CS & Sales sprint, and which seemed worth making an issue on for various teams to weigh in on. @charlescook-ph @raquelmsmith and @simfish85 especially may have thoughts here.
Proposal: Let's stop using Clearbit to gather information on users who sign-up with non-work domains.
Context
Currently, when a user signs up we ask them the following questions, both of which are optional:
We then ingest sign-ups into Hubspot (soon to be Salesforce) and enhance them with Clearbit to add the following information for ICP scoring:
We also use email provider and country as other fields which are gather automatically from login and IP.
We currently pay $15,000 per year to Clearbit for 75,000 data enrichments. This includes users who sign up with work and non-work domains.
Why change?
The main reason to change is in order to save money, which we'd do because the current enrichment system doesn't add a huge amount of value.
If a user signs up under a work email address, Clearbit enrichment makes total sense. We can assume that this user is signing up for that company in a professional context. Clearbit can fill in the gaps and help CS & Sales to spot big companies which they want to talk to.
However, if a user signs up with a non-work email address, enrichment makes less sense. We can assume they aren't signing up on behalf of the company Clearbit thinks they belong to, so it's unlikely CS & Sales will want to talk to them regardless of the enriched data -- a fact that is currently reflected in very strong ICP weighting against users on a personal domain.
Clearbit also isn't 100% accurate and we have no way to confidently validate the data. Best case scenario is that Clearbit correctly enriches a user on a non-work domain and that the data is mostly useless anyway. Worst case scenario is that Clearbit incorrectly enriches users on these domains and our ICP gets biased by bad data.
We currently score users on personal domains -10 in ICP scoring, but it's cheaper and more efficient to simply not enrich these users at all.
Why not change?
The only reason not to change this that I can tell is because we want to upsell users on personal domains into companies revealed by enrichment. However, we currently have no process for this.
There's also the slight chance that a perfect user would still come in on a personal email domain -- however that user would have to max out all other ICP scoring factors (a score of 24) and be based in an ICP target country to offset the -10 score from using a non-work email. That seems unlikely, or that it'd be based on Clearbit inaccuracies.
The text was updated successfully, but these errors were encountered: