Enriching HubSpot with Clay (Part 2: Names and job titles)

Welcome to part two in our five-part series on enriching HubSpot with Clay, focusing on outbound activities.

We start by revisiting our methodology from the first video, where we used LinkedIn profiles to identify CRM gaps.

In this video, we delve deeper into building lists in HubSpot and stitching data across multiple providers including Clay’s data, Apollo, and testing LeadMagic’s API.

We ensure accurate updates to HubSpot records using IDs. We’ve also introduced a GPT prompt to clean job titles, adding robustness to our data enrichment process.

In the next video, we’ll cover obtaining and validating phone numbers for effective outreach.

Transcript

Introduction to Series and Recap of Part One

[00:00:00] Harris Kenny: Welcome to part two in our five part series on how to enrich HubSpot with Clay, with a special emphasis on enabling outbound activities. But of course, I think you’ll find that the things that you can learn in these videos can be applied to inbound.

[00:00:13] Harris Kenny: I want to start with refreshing your memory about what we did in the last video. And if you haven’t seen it yet, I would encourage you to go back and watch it. In the first video, we established our methodology for how we’re using lists in HubSpot and then list sources in Clay to identify gaps in our CRM.

[00:00:32] Harris Kenny: The specific data that we found last time was LinkedIn profiles. And then we identified ones that were missing, found them, populated them. And moving forward, we’re going to use those LinkedIn profiles for the rest of the videos in this series. So if you haven’t done it yet, I really encourage you to watch that video or make sure that you have a LinkedIn profile property in HubSpot and that data in HubSpot so that you can use the rest of these.

Building and Managing Lists in HubSpot

[00:00:57] Harris Kenny: We’ve got our lists here on the left in HubSpot, and I’m just going to jump into how we built them here. You can see we’ve got either first name or last name is unknown. And in both cases, we’ve got LinkedIn profile is known. And the next one we’ve got job title is unknown, but LinkedIn profile is known.

[00:01:16] Harris Kenny: Now, the way that I’m doing this, there’s going to be potentially some overlap because there’s going to be examples of records where they have no name and no LinkedIn job title. I’m okay with that. I’m okay with paying a little bit extra here in order to have more control and granular control over how we’re doing the enrichments for each of these properties. So that’s a deliberate choice.

[00:01:37] Harris Kenny: You can also see that some of these are never going to enrich. For example, text@example.com, that will never have a contact name because it’s not a real account. There’s no real person there. It’s never going to resolve. I think a potential enhancement you could do to this would be to add a property in HubSpot to indicate that something is omitted from enrichment or it’s, it’s not to be enriched.

[00:01:59] Harris Kenny: And if you want to use the size of these contact lists to indicate the health of your data where, you know, you want them to be zero, uh, then I think that would be something that you might want to consider.

Exploring Clay Tables and Data Providers

[00:02:10] Harris Kenny: Let’s jump into the Clay tables because they are, I’m going to show you a few things that I didn’t show in the first video, but with greater detail, I think that will be useful moving forward, including a GPT prompt and how we use the IDs from HubSpot to make sure that we’re updating the right records.

[00:02:24] Harris Kenny: Looking at Clay, we’ve got our table. Once again, HubSpot is the source for this list. And in this case, we’re stitching across three different providers. Now, one of them is Clay’s own data. One of them is Apollo where we’ve got an API key. And then a third is a new vendor, uh, called LeadMagic. And we’re testing LeadMagic’s API here.

[00:02:44] Harris Kenny: Now there are interesting situations where you’re going to want to try different vendors, different providers, because you can find unique data points from them and they may not be supported in Clay. In a later video, I’m going to talk about how LeadMagic actually has a new API endpoint, and we’re going to introduce that new LeadMagic API endpoint into a waterfall.

[00:03:03] Harris Kenny: I’m not going to talk about that right now. I just want to show that the benefit here is configurability and how you can change over time the way that you’re doing this enrichment. You can reward more accurate, or less expensive vendors, or both. You have control basically over what you’re paying for and when and why.

[00:03:17] Harris Kenny: There may be special requirements you have for your industry. And that’s where this approach can be really helpful. Beyond knowing what’s going into your system, you can have a lot of control over how it works.

[00:03:26] Harris Kenny: Now on the left side, we’ve got that HubSpot source. We’ve got that ID.

Merging Data and Enriching HubSpot Contacts

[00:03:30] Harris Kenny: That ID is critical to make sure that we’re updating the right contact in HubSpot, once we get to the end of our table here on, on the right side of the Clay table.

[00:03:38] Harris Kenny: I’ve got a merge column that’s pulling across multiple different providers. And if it, if the first one finds it, we’re going to use that first name. If the second one finds it, we’re gonna use that one. Third one down the line. And then we’re going to take those name properties and push them over to HubSpot.

[00:03:53] Harris Kenny: This is what that merge looks like. This is what it looks like where it’s pulling across multiple different providers and then combining it into a single cell here. You can see the last names. In this column. And lastly here, what I’ve shown you is what the actual HubSpot enrichment looks like. Now this is an example, but here I really strongly recommend using that HubSpot contact ID instead of the contact email address.

[00:04:16] Harris Kenny: This will ensure that you’re getting the right contact updated. And in this case, We’re going to be updating their first and last name.

[00:04:23] Harris Kenny: Now the use case here beyond just having nice clean data is if you’re going to be incorporating omni channel outreach, cold calling and things like that, it can be very useful to have the contact name of the person you’re calling.

[00:04:34] Harris Kenny: And, you know, depending on where your sourcing leads from, you may not have it in the first place. This is going to be, I think, a really valuable sales enablement piece of data that, um, teams especially doing phones are going to want. But you can also use it for variable personalization. You may be getting inbound leads through email signups, through lead magnets on your website where you’re not getting this type of data.

[00:04:53] Harris Kenny: So there’s other situations where it’s useful. I think from a revenue perspective, cold calling is probably the most valuable, um, situation where you want to make sure you have at least contact first names.

[00:05:04] Harris Kenny: Okay.

Importance of Job Titles in Outbound Campaigns

[00:05:04] Harris Kenny: Now let’s talk about job title. The use case here I think is interesting because it’s really going to be valuable for future prospecting efforts as well as inbound and understanding your funnel and things like that.

[00:05:14] Harris Kenny: But again, with our outbound focus, you can take this list of job titles and you can give it to your outbound agency, you can give it to your own biz dev team or growth team and you can have them have a much better understanding of who they’re reaching out to.

[00:05:26] Harris Kenny: This is where going back to that HubSpot step where we built those filters in the first place.

[00:05:31] Harris Kenny: Really matters because you can set filter requirements for the life cycle stage or associated deals or associated with closed won deals or associated with deals of a lifetime value of over $10,000 or something like that. And you can look at segmentation and get a sense of who these contacts are in your system based on their job title.

[00:05:50] Harris Kenny: But if you do not have that job title data, you can’t do that type of segmentation, reporting, and you lose a feedback loop to inform your outbound campaigns.

[00:05:59] Harris Kenny: Most campaigns that fail, fail from the start because the list is the strategy. Often repeated. I think it’s absolutely true. So I’ll repeat it here too. The list is the strategy and you can use job titles to build better lists. But how do we do it?

[00:06:16] Harris Kenny: Once again, we’re stitching across multiple providers. We’re using LeadMagic. We’re using Apollo. We’re using Clay’s data. In both cases, I found that these providers were pretty good fit for what we were trying to do here.

[00:06:26] Harris Kenny: We’re still getting unstructured data back. We’re getting some goofy situations here where we’ve got, for example, things in parentheses and extra names and more descriptive names and things like that.

Using GPT for Job Title Enrichment

[00:06:37] Harris Kenny: So we’re going to be using a Clay prompt and I’m going to just slowly scroll through this so you can see the whole thing.

[00:06:41] Harris Kenny: You can pause the video. Check it out. We’ve got a generate text enrichment in Clay. We’ve got a couple of examples that we provided. We are requiring that a raw title be present in order to run this column, which is going to help save credits, because you don’t want to run it unless you’ve got something to feed into it.

[00:06:59] Harris Kenny: And then this is what we’re reading into chat GPT. First we’ve got our system prompt, which is providing context for the task overall, explaining what the GPT’s responsibility is.

[00:07:10] Harris Kenny: And then in the second one, we’ve got the specific task, but we’re explaining what to do, reminding it, what to do in cases where it’s not sure and giving some context about how extraneous information may be included.

[00:07:22] Harris Kenny: And finally, we’re providing it with the raw job title that we got from the providers earlier.

[00:07:27] Harris Kenny: What you can see here in this column where it says title ready, that’s the output. And we’re requiring, before updating HubSpot, that the title is not unknown. So this is a really useful way to make sure that you run this enrichment, clean up your data, but not introduce new data problems.

Conclusion and Preview of Next Video

[00:07:43] Harris Kenny: I hope you found this helpful, trying to drill down into a little more detail here. Talk about some potential use cases. In the next video, we’re going to talk about phone numbers. How to get phone numbers, how to stitch across multiple providers, and how to account for the fact that some providers are much more accurate than others.

[00:07:59] Harris Kenny: And what do you do when you only have maybe one good one, but a second number that might be good and you want to give your team options to stay on the phone and try to get in touch with this contact.

[00:08:09] Harris Kenny: This is something I’ve built out myself for our own prospecting and even inbound engagement efforts that I’ve found to be extremely useful.

[00:08:17] Harris Kenny: That’s going to be in the next video. We’re then going to talk about updating our waterfall once we discover new providers, and then finally, we’re going to end with tracking job changes from your CRM using Clay.