It was a post by Sara McNamara on LinkedIn which drew our attention to the fact that, to the extent account matching depends on tracking corporate IPs, it would run into problems if everyone started working from home. Which, of course, is largely what has happened over the last three to four months.
Is there an IP lookup problem?
If McNamara, senior marketing operations manager at Cloudera, and something of a thought-leader in the ABM and marketing automation spaces, detected a problem, it was worth asking some other practitioners.
One senior digital marketer, and a seasoned observer of ABM strategies, told me on background: “We are very focused on a list of defined accounts. When everyone started sending people home we were pretty concerned about what that would do to our ability to track and target the employees at those key accounts. About a month in we did some analysis and worked with our IP lookup vendor, and were able to evaluate what the impact of work from home was. In short there was an impact but it wasn’t material enough to affect our marketing in the April to June time-frame. As this goes longer the situation may change but as of now we’re monitoring but not actively worried.”
A critical part of a modern ABM strategy is crawling the web to identify activity which shows a user researching products and services in a way which signals they might be in-market, and then mapping that user to an account—often by matching the user IP address to a corporate IP address.
Since acquiring Spiderbook in 2016, Demandbase has been recognized as having a powerful, AI-driven tool for identifying in-market accounts at scale. CMO Peter Isaacson told us: “During the early days of the pandemic, there were a lot of questions about whether an ABM approach would still work, since it might be more difficult to match individuals to companies with so many people working from home. However, we were able to maintain, and recently, even increase our match rates and accuracy during this time.”
The key, Isaacson explained, has been combining proprietary company IP matching data with individual cookie data. “Together, this helps us correctly identify companies, even as employees are working from home, and target those accounts with ads or personalized web experiences.”
The cookie trail
Terminus is a smaller and newer ABM vendor than Demandbase, but sits alongside it as a leading vendor in the 2020 Q2 Forrester Wave for ABM. Last week it announced a significant platform enhancement with the introduction of the Terminus Engagement Hub, described in a release as “the only all-in-one full-funnel ABM platform.”
Among the new capabilities are an enhanced ABM scorecard which tracks trends in market segments over time; improved tools in the Data Studio to identify next-best-actions based on ad performance and engagement metrics; and an ability to import and connect Salesforce account IDs.
Justin Keller, VP of marketing at Terminus, and Daniel Hellerman, VP of digital ads, took us on a deep dive into the world of ABM cookies. We asked Keller if he was aware of the concerns caused by remote working.
“For sure,” he said “it’s been a big topic. We use IP targeting here, but we’ve also got a really deep cookie pool that we resolve so that we can follow people at target accounts wherever they are—at home, across the country, wherever.”
Hellerman explained the process. “There are a couple of different ways you supplement IPs with cookies. One is that you can overlay cookies with IPs. Say you buy a portion of cookies from [a vendor] like Oracle representing people in marketing.” Those can then be matched with IP addresses of interest. The second method is more relevant to the current situation: “You start with the IP address of the company to generate the cookie.” Those cookies then land on any device which ever accesses the corporate network. “You can then move that pool over to supplement existing data and overlay it: of the cookies I have.”
There’s even a workaround for people who clear their cookie cache. If they interact later with a corporate website from the same IP address, the new cookie can be matched with the historical cookie, even though the user has since cleared it. The user’s browser history can help confirm the identification. Vendors like Oracle, said Hellerman, can also match retail transaction data with historical records to improve accuracy.
Has Terminus changed its strategy in response to COVID? “We’ve been doing it in any case,” said Hellerman, “relying on Dun & Bradstreet and Oracle as our main cookie vendors. The data is continuously refreshed, with vendors using all these signals to do it. I would agree that it would be a lot harder to target really small companies that may not have an identifiable [corporate] IP in anybody’s database, but most of our customers are targeting the Fortune 10,000 or 20,000, which are very easy to identify.”
If people were not only working remotely, but changing jobs and roles during this time, the matching would be much more of a challenge, said Hellerman. But of course COVID has discouraged job-switching.
6Sense is another leader in the Forrester Wave for ABM. Last week, it extended its suite of advertising solutions with Native Retargeting, the capability of serving relevant ads across channels to target account representatives who had already visited a corporate website. “Our strategy is to provide customers access to the broadest network of advertising inventory, and then overlay our patented company identification and AI capabilities to ensure campaigns reach their ideal accounts,” said co-founder and CTO Viral Bajaria in a release.
We asked Bajaria about the impact of COVID on their company identification strategies. “When we started our graph [of companies], almost five or six years ago, we did put a focus on corporate IPs, but we didn’t rely on them completely. We are a Bay Area-based company, and Bay Area companies have always done a lot of work from home, so our graph and technology was really built for this.”
Probability and confidence
Like Terminus, 6Sense cookies people using the corporate IP address in order to re-identify them when they work remotely. “It’s not like we have a Facebook-like log in, so ours is a probabilistic graph. We do get false positives, but not a lot, and we have a confidence metric. “Are we highly confident that this visitor, although they’re not from the corporate IP, belongs to company A? We have high confidence, moderate or low. You can rely on that confidence to decide what you want to do. In some cases, where you’re personalizing stuff, you get rid of the low confidence.”
The shift to remote working caused by COVID did produce a drop in account-matching for 6Sense. “We did a re-analysis in July,” said Bajaria, “and we are now we are almost back at our original level.” 6Sense did undertake some small adjustments. “Previously you would not have had multiple people working from home. That wasn’t happening a lot,” Bajaria explained. “We had to make some changes to adapt to that, and change our confidence coding.” It’s possible, again probabilistically, to identify that two people—spouses, say—are working from the same IP address, but for different companies.
“When you are doing broad-based advertising, it’s okay to get some wrong. When you are hyper-focused and personalizing, you have to get it right, so you narrow your audience.”
Despite the fact that the value of corporate IP-tracking has been highly touted in the past in the ABM space, it seems that some vendors already had alternative strategies—initially to supplement the corporate graphs, and recently to provide essential support. According to McNamara, Cloudera, as a Pardot user, had been able to use Pardot cookies to address a Demandbase identification drop-off.
Third-party cookies, which Google is set to phase out in Chrome by 2022, are not really in play here, but the question remains: as ABM becomes ever more critical to B2B sales and marketing strategies, how could it survive an all-out war on cookies?
This story first appeared on Search Engine Land.