Finding Meaningful Patterns in Intent Data Requires Advanced Analytics
By Ray Estevez – Chief Data Officer, True Influence
In my recent post about how we here at True Influence collect intent data to fuel our demand generation and market intelligence products, I noted that the essence of Big Data is knowing how to correlate billions of pieces of information to find meaningful, actionable patterns.
This, of course, boils down to analytics, and our Relevance Engine™ analytics platform has been central to our success since 2010, when we launched it and helped define B2B intent monitoring as a martech category. Since then, we’ve constantly honed Relevance Engine’s algorithms to find patterns in purchase intent activity that go well beyond just recency and frequency.
We’ve also developed advanced, multi-factor Identity Graph analysis to help determine exactly who is exhibiting purchase intent, both at the account and individual level. And we employ natural language and other machine learning analytics to determine the real topics prospects are researching when they engage with content online.
In this post, I’ll explain how our advanced analytics work together to find meaningful, actionable purchase intent in all that data.
The True Influence Identity Graph
At the most basic level, an Identity Graph is a database that contains and cross-references all the known identifiers you can collect about a business or a contact. Signal.co has a nice primer on the topic that explains the wide array of information that can be included in an Identity Graph, such as physical location, device IDs and online behavior history.
At True Influence, our Identity Graph includes dozens of values that we use to identify exactly where an intent signal is coming from. We employ advanced multi-factor analysis to build connections between the intent data we gather and what we already know about the business or individuals in our massive B2B contact database.
Now’s a good time for a little background on the exact nature of the intent signals we collect from our sources. Some signals come with cookie data or other explicit, opt-in information that lets us know exactly who downloaded a whitepaper or attended a webinar. Other signals may have a hashed or encrypted email identifier. And others may have only an IP address.
In some cases, all we know is that someone has raised their hand and shown interest in a topic.
From this starting point, we cross-reference every piece of data we have to determine exactly who it was that raised that hand. For example, if we know that a signal came from a large company’s upstate New York campus, based on IP address, we’ll cross-reference that information with device IDs and other specific identifiers we have for that location, along with the date and time the event occurred. We’ll then query our database for contacts at that campus with job titles and functions that match the topics we’ve assigned to the intent signal (a little bit more on that later), then cross-reference these likely candidates with even more identifiers to determine if they are the individual who expressed intent.
Ultimately, this exhaustive cross-referencing results in us being able to target in-market contacts within an organization, based on the intensity of the purchase intent being exhibited.
This is critical in intent-based B2B sales and marketing. You just can’t blast the entire upstate New York campus of a large corporation with a generic, high-funnel marketing offer and expect positive results. Intent is about improving the efficiency and ROI on your sales and marketing efforts, and that’s not going to happen unless you know which individuals are the best candidates for outreach.
We use our Identity Graph to map intent to individuals for our demand generation programs and in our subscription products, including InsightBase®, which we expanded in version 4.0 to include buying group intelligence based on the ability to identify in-market individuals within organizations. Our Identity Graph also helps us identify smaller businesses and locations that don’t have dedicated, easily identifiable IP blocks associated with them.
It’s a powerful piece of technology that sets us apart.
True Influence Relevance Engine
Not every blog post view means that an account is primed for a call from Inside Sales.
That’s the wisdom behind Relevance Engine, our proprietary analytics platform that finds patterns of purchase research; determines how this activity maps to your products and services; and then spotlights individuals and accounts who are ready for marketing or sales outreach.
Relevance Engine looks at the last eight weeks of intent data to determine if research activity is on the rise – the clearest indicator that purchase intent is spiking. And, as I’ve said earlier, Relevance Engine bases its analysis on more than frequency and recency – it also considers firmographic intelligence, such as revenue, number of employees, and significant events in recent company history.
If a company just had an IPO, it’s likely looking to use that influx of new capital for investments hiring, capex infrastructure spending or development of new systems. Now’s the time to talk.
Relevance Engine takes all these factors into account and then calculates an intent ranking that indicates the next, best steps for connecting with an account or individual. A hot account may be ready to be passed directly to BDRs for phone outreach; a plateau in intent activity may indicate that a steady nurture campaign is in order.
Relevance Engine is critical in translating those billions of intent data points into a strategy that grows your business. Our engineers have put a tremendous amount of energy into creating and developing it, and we continue to perfect it.
Natural Language Content Analysis
Our purchase intent intelligence is based on topics. We analyse the online content B2B prospects engage with and map their topical interest to your product or solution.
In order to do that effectively, we need to understand the central themes of the content that prospects are reading. And that goes beyond simple keyword counts.
When we know the URL of a content item, such as a blog post or news item, we run a natural language processing (NLP) scan to find patterns that indicate exactly what the article is about. This advanced analysis goes well beyond simply counting the occurrences of phrases, checking for terms in subheads, and other simplistic checks employed by many generic content scanners.
We also include SEO keywords included by webmasters (we find that this data is implemented and useful in about 80 percent of cases), IAB advertising data targeting tags, and other general topic-related data in our content analysis.
The net result is that we typically identity three to five main topics that a content item addresses, then catalog the intent signals associated with it in our 6,000-term B2B taxonomy. Our content analysis technology also allows us to add new topics and find meaningful intent around them within a matter of days.
The net result is that we identify contacts who are genuinely interested in what you’re selling.
It’s All About The Connections
As you can see, we never stop improving the analytical tools we use to find actionable sales and marketing intelligence in the billions of intent signals we collect and analyze. By correlating all that data, we give the market insights you need to reach the right contacts and close more business.