HubSpot Insights Database enriches wrong information (city, state, address line, ...)

Smart CRM & Sales Hub CRM
, , , , ,
1

Problem:

We have alot of German and European customers and therefore contacts. For some reason, a lof of our companies get invalid data enriched in the field, states/region, cities, address line, eventhough the domains are correct. I’ve found at least 800 cases, where the address information was enriched with invalid information. Also the employee number and revenue are wrong often times. It also happes for very big DAX30 companies, where there should be enough valid data. And in some cases Insights even added an American State and Postal Code, eventhough there was a German City and address line filled, which make no sense at all.

One of the reason, we chose HubSpot was to enrich our leads and to improve data quality, s

o disabling the feature is not an option for us and is not a proper solution to the problem. In fact, currently this reduces data quality for us and creates mistrust in data by our colleagues.

Assumption of root cause:

I have found a very stange pattern here: All invalid addresses are addresses of popular social network companies, that have been enriched by HubSpot to other companies:

- The Address of Google in 577 companies: 1600 Amphitheatre Parkway, Mountain View, CA 94043, USA

- The Address of YouTube in 119 companies: 901 Cherry Ave., San Bruno CA 94066, USA

- The Address of Instagram in 25 companies: 1601 Willow Road, Menlo Park, CA 94025, USA

- The Address of LinkedIn in 17 companies: 2029 Stierlin Court, Mountain View, CA 94043, USA

- The Address of Facebook in 2 companies: 1601 S. California Ave, Palo Alto, CA 94304, USA

- The Address of Twitter in 2 companies: 1355 Market St, Suite 900, San Francisco, CA 94103, USA

- The Address of Pinterest in 1 companies: 808 Brannan Street, San Francisco, CA 94103, USA

- The Address of AddThis in 1 companies: 8000 Westpark Drive, Suite 625, McLean, VA 2210, USA

Note: Not sure if there are more corrupt addresses by other companies.

All these addresses are in the imprint of most German and European company websites. One example can be seen here: Datenschutz | Liebich & Partner

Is almost seems like you might scrape this information from the companies pages and populate the data without validating.

Expected Result

  1. The data enriched by HubSpot is correct and valid. As seen in the patterns above, the data is not a 100% valid.
  2. HubSpot Insight should only enrich data, when the algorithm can make sure, that it is a exact match. Otherwise, the enriched data is wrong and creates more harm than good. If there is no exact match
  3. Also, there should be an option to enable Insights on a field basis. See here: https://community.hubspot.com/t5/CRM/Disable-Enable-HubSpot-Insights-for-each-field-seperately/m-p/419155#M125702

Update from the lovely Chiara from Support:
We had the issue of HubSpot Data Insights a few months ago in this ticket. The tool had incorrect date introduced into the CRM, and you had identified that multiple times the data was introduced from social media networks. The team has now updated and I wanted to let you know what has changed:

  • We have corrected the data on our site for the names and addresses, but it will not show up in your account until HubSpot Data Insights is re-enabled in the account, thus resyncing the data with the database.
  • Blocking known incorrect addresses and names from being included in the enterprise dataset unless they are on an approved exception list (e.g. 1600 Amphitheatre Parkway is the valid address for google.com, so we have set that as an exception to be properly included). Again, this is an ongoing process as we review potentially problematic names and addresses each month and add them to this blocklist where possible.
  • We take the errors we find to our third party vendors, have the data corrected on their end, and make sure they have improved their data collection process to reduce the number of incorrect addresses.
  • Fixing an error that caused the number of employees and amount of revenue to be displayed too high (trillions of dollars/employee). Unfortunately, this data is less reliable overall than things like names and addresses, especially for private companies, but we should now at least be showing the correct magnitude.

For this change to be realized in your account, you would need to enable HubSpot Data Insights here. If you prefer to work without this data and don’t want to enable the setting, feel free to give me some feedback on that. Due to the complexity of HubSpot Insights, the team had to spend more time on it and will continue to work on optimizing the tool.

Similar issues:

https://community.hubspot.com/t5/CRM/Incorrect-Autofilled-Data-Insights-for-Company-Properties/m-p/32642#M3541

https://community.hubspot.com/t5/HubSpot-Ideas/Filter-Rollback-of-all-HubSpot-Insight-changes/idi-p/358144

https://community.hubspot.com/t5/HubSpot-Ideas/Option-to-turn-off-HubSpot-Insights-on-company-records/idi-p/178713/page/2#comments

2

Hey @franksteiner79 & @karstenkoehler (or @Anonymous ?),

Ich nehme an eurem Datenqualität Webinar teil und denke, dass ihr dieses Problem kennt und auch wichtig finden könntet. Ist vorallem für deutsche Kunden echt schlimm, da es die Lead Enrichment Phase extrem stört. Freue mich über Feedback und upvotes. :slightly_smiling_face:

Mario von //SEIBERT/MEDIA

3

Hey there,

I found a big issue in you CRM-module that actively decreases data quality for European and German customers. I think a fix would be huge improvement for the adoption in Germany and Europe and would make the tools so much more useful. Maybe you’ll consider upvoting and using your influence to get this bug seen and fixed. :slightly_smiling_face:

Wish you all the best and have a great day,

Mario from //SEIBERT/MEDIA (Enterprise User)

@xenia , @ChiaraGaviraghi , @leoo , @LouiseBernstein , @EwoutM , @michaelwalton , @vishnu , @chrishogan , @Alberto , @roisinkirby , @Dylan
PS: I found your profiles through linkedin and you seemed like the right people to talk to. Sorry for those I forgot to mention.

4

HubSpot fixed the issue and answered this:

We had the issue of HubSpot Data Insights a few months ago in this ticket. The tool had incorrect date introduced into the CRM, and you had identified that multiple times the data was introduced from social media networks. The team has now updated and I wanted to let you know what has changed:

  • We have corrected the data on our site for the names and addresses, but it will not show up in your account until HubSpot Data Insights is re-enabled in the account, thus resyncing the data with the database.
  • Blocking known incorrect addresses and names from being included in the enterprise dataset unless they are on an approved exception list (e.g. 1600 Amphitheatre Parkway is the valid address for google.com, so we have set that as an exception to be properly included). Again, this is an ongoing process as we review potentially problematic names and addresses each month and add them to this blocklist where possible.
  • We take the errors we find to our third party vendors, have the data corrected on their end, and make sure they have improved their data collection process to reduce the number of incorrect addresses.
  • Fixing an error that caused the number of employees and amount of revenue to be displayed too high (trillions of dollars/employee). Unfortunately, this data is less reliable overall than things like names and addresses, especially for private companies, but we should now at least be showing the correct magnitude.

For this change to be realized in your account, you would need to enable HubSpot Data Insights here. If you prefer to work without this data and don’t want to enable the setting, feel free to give me some feedback on that. Due to the complexity of HubSpot Insights, the team had to spend more time on it and will continue to work on optimizing the tool

5

Thank you @CNicholl for letting the team know. One last question. Could you maybe ask them, if they could simply make it possible to enable/disable Insights on individual fields e.g. enrich Company Name but not Address. In my opinion, this should be a easy fix (1 day) and would give the user more control over what is happening in HubSpot. Then this feature is perfect.

6

We’re using data enrichment very selectively to enrich company records and I found that it still makes too many mistakes, even though they could be easily avoidable.

A small company with a domain name ending of .ch (a company from Zurich, Switzerland) had the city name “New York” (but interestingly not the country USA) and another European company with a .de (German) domain name and no subsidiaries or activity in the US had the city name “BROOKLYN” (in capital letters!) set by the enrichment tool.

It’s obviuos that most of these errors could have been avoided if the AI simply visits and parses the websites of these companies and parses their content. I’ve manually checked some of the faulty data and I found the correct addresses and data in a matter of minutes on the respective company website.

At this point it’s more of a “nice to have” add-on rather than something trustworthy.

7

Hi @MKnoedgen, I’d like to thank you for sharing your valuable feedback! This means a lot to us.
I have shared it internally to bring more visibility to it.
I see that you commented on an old thread, so just curious, are you using the new AI data enrichment?
Please let me know if there is anything else I can help with. I’ll be delighted to do so!
Have a great day!
Bérangère

8

Hi @BérangèreL – yes, sorry if this was confusing. I’m talking about the new data enrichment feature you can now use for companies and contacts. I would love to use it to improve general data quality but I think it’s not yet viable for a daily use (at least outside the US).

9

Thank you for sharing your experience with the new data enrichment feature @MKnoedgen.
Your input about the current limitations outside the US is very valuable, and I’ll make sure to pass along your feedback to the product team.
We appreciate your insights as we continue working to improve data quality for all regions!
Thanks again, and please let me know if there is anything I can do to help you, I’d be delighted to do so!
Bérangère

10

Hello MKnoedgen ,

I’m sorry to hear about what happened to you. My company has been using Enrich-CRM for several months, and it has been great for data enrichment. We integrated it with our HubSpot, and it enriches data live while providing accurate information.

From my experience, why not give it a try? It has met all of our needs so far. Since we started using this tool, we haven’t faced any issues like the one you mentioned.

Highly recommended!

Have a great day,

Lamy Stevenson