How do I get Glassdoor ratings for a set of companies?

Hi there,

Been playing about with Rows :clap: so much potential

I’m running some analysis on a set of companies by domain. In the absence of a Glassdoor integration, how do I get the Glassdoor ratings for each company? For now, I’m after the headline star rating, but there are other data points I’d like to retrieve.

I’m assuming here Glassdoor is not on your integration roadmap in the short term…?

Glassdoor API Access

From what I understand, I could use the GET function and my own API access to do this, but I’m currently waiting on Glassdoor to respond to my request.

In the meantime, I’ve got a few questions here:

  1. While I wait on API access, is there a way to get at least the Glassdoor star rating via the Google Search Console API. The overall star rating almost always shows up as the first Google search results—albeit there sometimes a mismatch between the results based on whether it’s from Glassdoors company /overview page or /reviews. PS. I tried this already, but it didn’t seem to pull the rating score at all.

  2. …Or as an alternative, are such Glassdoor data points contained within any of your existing integrations?

  3. And let’s say a company has multiple domains, e.g. Intercom has .com and .io and probably some others; how do I ensure it looks up against one or the other(s), or both/all*.

For example, for Glassdoor, Intercom’s rating is tied to the .com, yet for Trustpilot, 99% of their ratings are attached to their .io domain.

I’d want some logic to pull what’s most relevant based on volume or some sort of concatenation calculation where the ratings are split across more than one than domain (hence the * against both/all above).

Regarding the Glassdoor API, I’ve no idea how expansive it is and not sure what I’ve requested access to exactly, and I suspect my request may be in a contact us vortex :rofl: Apparently, they’ve public and private API, but seemingly you need to be an API partner for both. If anyone can shed any light on that, that would be great! Surprisingly little I can find on this. A lot of people seem to resort to scraping.

Anyway, API noob here and Glassdoors API documentation is lacking!

1 Like

Hey @lynnastyie :wave:t4: — thanks for reaching out and it’s great to see how well you’ve been exploring the platform!

You’re right about Glassdoor not being in our immediate/short term roadmap; nonetheless, I’ve added it to our list.

It appears that their free API sign up is currently closed -

I too tried retrieving the ratings via our Google Search - Top Sites integration (I suspect this is what you may’ve meant and not the actual Google Search Console integration) but to be fair, working with Google’s programmable search is more of an art than an exact science.
Fortunately, there were some indications of the rating data that I was able to observe -

In my case, here’s the configuration I used when creating a Search engine ID via -

Feel free to tweak it as per your needs.

I suspect you may’ve already tried it but if not, we do offer Trustpilot as part of our instant integrations :muscle:t4:
When using Trustpilot though, there are no wildcards that you could use to look-up both .com and .io related reviews in a single function execution. To do that, you’d need to run the function twice — once on the .com domain name and another on the .io one.
We have no way of knowing what other domain extensions are being used by a specific company.

I’m afraid we currently don’t have Glassdoor data as part of any other integrations either but once you’re able to obtain the required API credentials / access, do please feel free to reach out again and I’d be happy to help you build that workflow.

Hope these help!

Hey @sourabh :wave: thanks for that info. I hadn’t seen that Glassdoor page about the free API being closed. That’s a shame. Be interesting to see what they come back.

Google Search Console

Right, got you. It was actually the Google Search Console I used. I picked up that tip from one of the Webinar video recordings over on Youtube. That was possibly from before the integration with LinkedIn (as the webinar mentioned circumventing going through LinkedIn’s API).

I’ve been able to replicate your configuration and skip through the results (per the tip mentioned in that webinar at this specific point). It’s one of those things where for this use case, it sort of works with a handful of records but it’s very flaky :laughing: I’m feeding in variables along with the company name to make sure I get the right company e.g. Zoom Video (in a half baked attempt to avoid the wrong company, say Zoom Info). Marginally better than looking them up manually.

I managed to pull through “CEO approval rating” as a snippet but couldn’t find a formula to extract the % value only from the snippet (Is there one?).

Similarly, I needed to implement a results Id reference in another cell to be able to pull through the most appropriate snippet i.e. from the right company and something dated recently. Using the console route is very much an art as you say!!

Overall, my approach here is not very scalable. Even if I got it working, with search results moving around, it breaks and there’s a high risk of parsing the wrong company record and/or something that’s out of date on the snippet front.

Glassdoor Nuances

Glassdoor also has reviews for some companies split by office GEO. But from what I can tell they’re still grouped under an overall employer record. It’s hard to tell when going via the Google search console unless you’re in the browser. I expect the API makes it easy to retrieve the overall one. But some companies seem to deliberately split their Glassdoor ratings (international level it seems). E.g my former employer has .com Glassdoor score, it’s US-centric (and for clarity I mean the record is listed as a separate employer), whereas the is UK-centric (albeit I’m sure it’s a mix of the two GEOs (given job listings are for both countries). If I get API access, may just have to ignore that unless I can solve the associated domains/domain extension issue and group them under one record.

Overall, I’d be better with the API. At least I should be able to get the overall rating for a given domain. But the source data will have other nuances and quirks. Age-old data problems!.

Anyway, the Google Search Console route was worth a shot. Thanks for your help there.

Associated domains / Domain extensions

I was going to ask in my first post whether finding extensions was doable. I only spotted the Intercom example because in another test query with the Trustpilot integration it only returned 2 ratings based on the .com. I knew they had the .io and sure enough, it returned something more plausible rating count wise.

You can probably guess where I’m going with this, but the domain (domain path? is that the right term) is my key identifier for a company. In an effort to close some gaps, I was looking for an automated way of finding associated domains that I could then build up from. But, you’ve answered the question.

I’m surprised none of the company data aggregators do this in an effort to clean their data. There are definitely a lot of duplicate records (on Dealroom etc.) but I guess it’s hard to distinguish whether it’s the same “company”, a subsidiary or something else (without human intervention and the company themselves feeding into that).

Duplicate records aside, it’s a pity none of them lists alternative domains. A possible workaround that I’m thinking of that would probably cover 90% of cases is hooking into multiple APIs based on a parameter other than website domain (Twitter account URL, Linkedin URL) and then retrieving the website domains and doing an automated cross-reference. And then go from there.

Case in point Crunchbase lists Intercom’s .com domain while Dealroom lists the .io but they share other parameters in common.

Company Domain Search

Somewhat related to that (didn’t seem like a separate thread was worthwhile albeit this response is long!!). For company domain search, how do I know I’ve got the right one? Is it a case of manually checking, checking against Crunchbase?

Is there a function where it can return more than one response? Take Clubhouse as an example, there’s clubhouse .io (the project management software) and joinclubhouse .com (the Clubhouse suffering a bit of a :chart_with_downwards_trend:). Similarly, there’s Bolt, one is the online checkout company (.com), the other is mobility (.eu) …and sure there might be more Bolts. If there isn’t such a function to retrieve more than one response, how does the Find Domain function choose which one to retrieve? Does it pick the first one based on Google ranking for the company name?

Haha, I think I’m really wishing someone would have created a company equivalent of the ISBNs they use for books, cause companies rebrand too and change domains like dashdash to Rows :grin:

EDIT: I can only post 2 links as a new user so had to leave some spaces in some domain refs above.

1 Like

Thanks for sharing such detailed notes, @lynnastyie :raised_hands:t4:

Let me try and answer some of the questions you’ve asked -

  1. formula to extract the % value only

I suspect this should be possible using our PARSE function although, it might help for me to actually look at the data / JSON and then be able to help identify the best approach here — I’m happy to jump on a quick screen share session too so we could go over it together.
Please feel free to DM the deets :blush:

  1. The right domain from the FIND_DOMAIN function

We try and do a best-effort match that’s closest to the input value. This is not always the first one though. There’s also an internal algorithm that tries to ascertain which extension ought to be given preference but you’re right — given that the approach here could lead to incorrect results, the best way to know if it’s the right one would be to check manually, use our Crunchbase integration functions or any other API that closely provides more accurate results.

  1. more than one response

You can use our GET_COMPANY_DETAIL_CRUNCHBASE function to find companies/domains that match the input value; however, I’m unsure how much of it would really solve the problem that you’re describing -

In my experience, I’ve found that content curation — be it tweets, YouTube videos, company data / domains — can only be automated so much and to really get what you need, some amount of manual intervention would be required.
I :heart: your idea about ISBN equivalents for companies too — if you do find an API that does this, do please let us know :wink: