Cleaning up contact duplicates in a regulated industry
Karin Tamir (00:00):
We specialize in support and complex integrations with HubSpot.
Jonas De Mets (00:06):
Nice. And in those projects, do you often run into duplicates? Any specific use case you'd like to highlight?
Karin Tamir (00:16):
Absolutely. One of the things that drew my attention to Koalify was the fact that in every HubSpot portal we work with, duplicates are always an issue. This usually happens because most companies today use enrichment tools—or even if they don’t, they often import messy data.
The use case I want to share today is quite unique. It doesn't come from enrichment tools or standard imports but rather from an integration with a regulated credit line platform. This company offers non-bank credit lines through its own platform, which is integrated with HubSpot.
Because it’s a very regulated industry, data accuracy is crucial. And yet, the integration was creating tons of duplicates—so many, in fact, that I can’t even remember the total number before we started using Koalify.
Jonas De Mets (01:52):
Wow.
Karin Tamir (02:04):
Before Koalify, their project manager—who’s also the CMO—had to manually merge duplicates every morning. I built a dashboard for her that showed duplicate records, and she’d go through them one by one.
Jonas De Mets (02:24):
That sounds exhausting.
Karin Tamir (02:34):
It really was. The problem was that people would submit forms without emails or with incorrect ones. Later in the process, the platform would create a new contact record—sometimes without matching data—and that's where the duplicates started.
Jonas De Mets (03:09):
So if I understand correctly, the duplicates mostly came from early funnel form submissions via HubSpot, and then again later when they entered the platform and submitted new data?
Karin Tamir (03:28):
Exactly. It's not a reconversion—it’s one continuous process. A user starts with a HubSpot form, then moves to an external screen that's part of the platform. They go through authentication, upload documents, and only provide an email late in the process. Since HubSpot doesn’t use phone numbers as unique identifiers, duplicates stack up.
Jonas De Mets (04:21):
Got it. What were the biggest challenges caused by these duplicates?
Karin Tamir (04:30):
We had several. One was identical phone numbers with slightly different names, which HubSpot doesn’t flag. Another common case was same name, different emails. But overall, HubSpot doesn’t support phone numbers or custom properties as unique identifiers, which made it hard to prevent duplicates.
Jonas De Mets (05:11):
Did this impact the customer experience?
Karin Tamir (05:23):
Somewhat. The platform handles most of the customer communication, but internally, the problem was huge. We’d see three or four duplicate deals for the same contact. It wasn’t just confusing—it messed up reporting and operations.
Jonas De Mets (06:26):
So how did you solve it? You mentioned a dashboard?
Karin Tamir (06:36):
Yes. I created a health dashboard—something I do for all clients. It included key metrics, and for this client, duplicate contacts were front and center. When I started, they had thousands. The CMO would manually go in and merge them every day.
We worked together—along with your help—to create merging rules that matched how she manually merged. It was hard at first for her to trust an automated tool, but we replicated her logic using custom properties and rules.
Jonas De Mets (08:51):
Makes sense.
Karin Tamir (08:58):
We now rely on mobile numbers to detect duplicates. We created rules that mimic her manual process: first, check if the contact came from a specific form; then look for internal emails; next, evaluate free email providers like Gmail or Hotmail; and finally, use creation date if needed.
The workflow currently focuses only on mobile numbers. We add a delay to allow internal processes to run, then perform the merge. And it’s made a huge difference.
Jonas De Mets (10:47):
Nice.
Karin Tamir (10:52):
Of course, some duplicates still exist—we’re only targeting certain ones right now. But mobile is key, especially since their authentication process relies heavily on it. Without a valid mobile number, users can’t proceed.
Jonas De Mets (11:31):
So once a duplicate is detected based on the mobile number, your workflow identifies the primary record based on those custom rules, then merges it?
Karin Tamir (11:45):
Exactly. And now, it’s all automated.
Jonas De Mets (12:10):
That must save a ton of time.
Karin Tamir (12:13):
It really does. We applied a similar process to companies as well. For them, "company" doesn’t mean a domain name—it’s a unique identifier from their internal platform. So we built separate rules there too, again with your support.
For this client, it's more than just data hygiene. Their entire industry is regulated, and at the end of the year, they need to report exactly how many loans were issued and how much was withdrawn. If their data isn’t accurate, they can’t do that.
Jonas De Mets (13:35):
Makes perfect sense.
Karin Tamir (13:51):
We now monitor their data daily. And honestly, Koalify has been a lifesaver.
Jonas De Mets (13:59):
Thanks for that. Any rough idea how many contacts were merged?
Karin Tamir (14:08):
Let me check... So since we launched the workflow—around October 30—we’ve merged around 1,500 contacts.
Jonas De Mets (14:43):
That’s a solid number for just a month and a half!
Karin Tamir (14:47):
Definitely. It’s a big deal for them. Now that she sees it works, she’s totally on board. I wish we could do the same for duplicate deals too!
Jonas De Mets (15:11):
That’s on the roadmap! Might even be released by the time this episode goes live. So, what are your biggest takeaways?
Karin Tamir (15:20):
I now use Koalify for every ongoing client we work with. Maybe not for brand-new portals, but for any client who’s been on HubSpot for over a year—it’s one of the first steps I take.
Showing them how many duplicates they have and what Koalify can clean up gives me an edge. In fact, I don’t know any other HubSpot partner in Israel using it right now. It’s a clear value-add.
Another lesson: always be specific in your rules. Avoid vague conditions like “when known”—it can cause real messes. We’ve made those mistakes and fixed them, thanks to your amazing support.
Jonas De Mets (18:03):
That’s great advice. Merges are currently irreversible, but HubSpot is planning updates in 2025. They’ll retain old record IDs after a merge, which might open the door to unmerging in the future.
Karin Tamir (18:37):
That would be great. We’ve definitely felt the pain when a merge went wrong.
Jonas De Mets (18:57):
Same here. But I’m hopeful 2025 will bring that flexibility.
Anyway, thanks for your time and all the great insights, Karin. This was a fantastic kickoff for our How I Fixed Your Data series.
Karin Tamir (19:12):
Thank you for having me!