In a recent life, I used to joke that I was in charge of all the “creepy stuff” at a Marketing Technology startup.
I was a product manager in charge of things like identifying anonymous visitors, intent (more on this soon) and advertising targeting.
I helped build systems which would take an ip address, triangulate it between different sources of identification and settle on what company or account to attribute it to in a world at the time where most everybody was working from home.
Along the way, I talked with lots of providers of third-party identification and intent about what they had and how they do things so we could integrate it into our own systems.
I’m not in that world anymore, and don’t particularly care to get back into it, so let’s talk a bit about how it all works. And why this dude is about to have a bad time - even if he’s right about a lot of it!
So, for those not deep in this space, let’s talk about marketing.
Marketers at Business To Business (B2B) startups - newish companies, largely fed by venture funding, and largely selling to one another - are increasingly tied to sales teams and revenue numbers as part of an overall Go To Market (GTM) strategy (this space loves their acronyms).
As part of that, sales hunger for leads - folks to talk to about buying their software or services. And sales drives revenue, and revenue drives venture funding and so forth. When it’s working well, it’s all a very fun and somewhat virtuous cycle.
So a lot of marketing has become sales assistance and support as a primary driver. There are still some folks doing squishier “brand awareness” marketing but they’re few and far between.
Marketing works to drive folks to their website and offerings through content, or advertising, or social media posting.
Once those visitors land on the site, in order to turn them into sales leads, they need to be identified.
One way to do this is through what’s called “gated” content - things like ebooks, reports, templates and such that require a visitor to leave an email address and various other data to get access to the content.
The other way is to try and tease out who the visitor is using things like third-party cookies or their ip address to try and unmask them to one degree or another.
Then you can reach out to these companies to try and set up a proper sales call.
In addition to these first-party sources of leads, there also exists a universe of third-party “intent” offerings that can be used to build lists for sales teams to try and do cold outreach to, or to power Account Based Marketing (ABM) programs of advertising.
These platforms - Bombora, 6Sense, Demandbase, etc - do the same anonymous identification I mentioned earlier, trying to tie an anonymous visitor to a company. Then they look at the pages or sites they’re visiting and attach that visit to various terms or keywords.
And they can also tie the company identification to lots of sources of company firmographic (location, size, revenue, etc) data to filter lists of companies by.
So, they can make it so you can say things like “Show me all the companies with revenue above $50 Million in San Fransisco that have looked at pages about content distribution” and then you can hand that list of companies to your cold-calling/emailing team (typically called Sales or Business Development Representatives) or use them for advertising targeting using platforms like Terminus.
Many of these intent platforms will also calculate things like “spikes” or “surges” which are intended to indicate a change in the consumption of things about these topics.
The problems with intent
Now that we have the background out of the way, here’s some of the ways this can go wrong.
A lot of the data sources behind this are shared between providers. They’re all essentially using the same blocks of ip addresses and the same wells of third-party cookies and drawing intent from the same set of media partners and the like.
For the company identification, they’re making guesses based on various probabilistic algorithms across all these different data sources. Sometimes these guesses are really good! Sometimes, they are very wrong.
These guesses got a lot worse once the pandemic hit and everybody started working from home and were no longer tied to corporate networks in many cases. Additionally, a lot of internet usage is done through personal email logins rather than business email logins, so many of these sources have your gmail address and not your bigco.com email email address. These platforms and providers have to build bridges between those.
That’s before you get into coffee shops and working spaces and VPNs that muddy the IP data even more.
Finally, the third-party cookie data is also increasingly shrinking in relevance. Those cookies only work in Google Chrome these days and won’t for much longer.
When you’re targeting advertising off of this data, it’s fine. You’re wasting some of your advertising budget, but nobody gets all that upset about it. It’s weird when you see a LinkedIn advertisement for something that isn’t a great fit for you, but most folks don’t think much about it.
The problem you run into - and many of these platforms are starting to run into this problem now as many of their customers look for software subscriptions that they can cut - is when you hand this same data over to your sales team.
Once your sales team is calling or emailing folks cold from them visiting your site or their showing some level of “intent” about some topic or another, that can be a different story. One or two misses of identification and many sales users of these platforms write off the entire endeavor.
If you clicked through that LinkedIn post earlier, he mentions that he’s talked to customers of these platforms and didn’t find anybody all that excited about them. That’s largely feedback from sales teams which have likely been burned repeatedly by either bad account identification or flaky levels of “intent” being shown (where one or two page visits might be enough to “spike”).
As a final bit of foreshadowing, this is all hazy and sketchy at the company level of the data. I’ve looked at them and spot-checked a lot of them - they’re bad and inaccurate about the company identification everywhere. And if they’re not, they’re extremely limited to company-owned IP addresses and sure-fire identities meaning you have lots of holes and not a lot of useful data to work with - quality, but no quantity.
When I cross-compared many of these services for company identification against things like internal signals, the best match rates I found was less than a coin flip for accuracy. Maybe they’ve gotten better in the last couple of years - but the world’s been making it harder for them, not easier.
That linked LinkedIn post earlier also mentions launching a new thing that doesn’t just identify the company making a visit or showing intent, but the individual doing so.
I’ve talked with a lot of lawyers in my previous role … and that’s not going to fly if you’re doing any business or seeing any visitors from Europe - or California.
Privacy laws such as GDPR and CCPA mean you’re taking on a lot of legal liability when you start tying together personally identifiable information (PII). Even the company level identification we’ve been talking about so far often involves sweet-talking lawyers to get things across the line.
If you’re going to go down the PII rabbit hole and try to uncover the precise individual doing activities, you’re inviting a boatload of lawyerly attention the second you get big enough for them to care.
Not that I think you would!
Because, as mentioned earlier - the folks who want the individual level are sales users. Everybody wants the Glengarry leads, after all.
Today’s company-level intent providers are already struggling with coverage, misidentification, and eroding trust.
Now imagine you try to figure out who exactly at those companies are making this activity. And imagine some Sales Development Representative somewhere fires off an AI-generated cold email outreach or picks up the phone and gives them a ring before they get ahold of somebody that doesn’t match.
How long do you think they’ll keep doing that?
How many times do they get wrong numbers, angry people, and coffeeshop strangers before they stop using it entirely?
From my experience, that number is quite low … and my educated guess is any service trying to dig out individuals is either going to trigger that experience frequently - or, if they lean into ironclad identification, there’s not enough meat on the bone to be valuable.
And that’s all before the lawyers get hold of you!
Godspeed, my dude, and good luck - you’ll need heaping amounts of both.
Wait. So, what should I do?
If you’re a consumer like me and want to opt out of this world as best you can, use something other than Chrome, turn off third-party cookies, run a VPN and laugh.
If you’re a marketer using one of these platforms, check with your sales teams and see how they’re finding the accuracy suits them. If they’re happy with what these platforms are unearthing, keep going! I’m not suggesting you stop doing stuff you and your team think are working!
If they’re not happy, though, it might be time to try a different approach to marketing. Identify your audience, talk to them, be curious, try to be helpful! Don’t go looking for “intent”, look for ways you can be helpful and honestly help people. And then marry that with intent sources that are in your control - gated form fills, newsletter subscribers, social media followers (well, less this bucket), that sort of thing.
But that’s me. Typing this behind a VPN …
Stay safe out there, folks!