Intent Data Basics: Where It Comes From, What It’s Good For, What To Test

3
373

Share on LinkedIn

Intent data is a marketer’s dream come true: rather than advertising to mass audiences in the hope of getting a handful of active buyers to identify themselves, just buy a list of those buyers and talk to them directly. It lops a whole layer off the top of the funnel and finally lets you discard the wasted half of your advertising.

But intent data is a complicated topic. It comes from different places, has different degrees of accuracy and coverage, and can be used in many different ways. Here’s a little primer on the topic.

What is intent data? It’s data that tells you an addressable individual is interested in your product. Ideally, that individual will be identifiable, meaning you have an email address, postal address, device ID, mobile app registration, phone number, or other piece of information that tells you who they are and lets you communicate with them directly. But sometimes you may only have an anonymous cookie or segment identifier that lets you reach them through only one channel and not by name.

Where does it come from? Lots of places. Behavior your company captures by itself is first-party intent data.  This is probably the most reliable but only applies to people you already know. Behavior captured by others is third-party intent data.  It's the most interesting because it provides new names or new information about names already in your database. Search queries are an obvious indicator of intent, but search engine vendors don’t sell lists based on query terms because they’d rather sell you ads based on those terms. So most third-party intent data is based on visits to Web pages whose content attracts prospective buyers of specific products. A smaller portion comes from other behaviors such as downloads, email clicks, or social media posts.

How is it sold? There are two primary formats. The first is ads served to people who have shown intent; this is generally called retargeting. The people are identified by cookies or device IDs and can be shown ads on any Web site that is part of the retargeter’s network. The original action could happen on your own Web site – the classic example is an abandoned shopping cart – or on some other site. Intent-based ads on social networks work roughly the same way, although the users are known individuals because they had to sign in. Retargeting is arguably a form of advertising and therefore not intent “data” at all. But I’m including it here because so many retargeting vendors describe their products in terms of intent. The second format is lists of email addresses. These are clearly data.  The addresses are usually gathered through user registration with the Web site.  Alternatively, the email can be derived from the cookie or device through matching services like Acxiom LiveRamp.

How reliable is it? Good question. Hopefully everyone reading knows that data isn’t perfect. But intent data is especially dicey because it comes from many different sources, some of which may be stronger indicators of intent than others. Users generally can’t see the original source. In addition, the data is aggregated by assigning the original Web content to standardized categories.  These may not precisely match your own products, or they may be grouped together so that some highly relevant intent is mixed with a lot of less relevant intent.  These problems are especially acute for B2B marketers, whose products may have a very narrow focus.  There’s also an issue of freshness: intent can change rapidly, either because someone already made their purchase or because their interests have shifted. So behavior that isn’t gathered and processed quickly may be obsolete by the time it reaches the marketer.

How complete is it? It’s worth distinguishing completeness from reliability because completeness is a big problem on its own. Intent vendors won’t necessarily capture every person in market for a particular product. In fact, depending on the situation, they may capture just a small fraction. Some people may not visit any site in the aggregator’s network; some may not visit often enough to register the required level of interest; some may decline to provide their identity or delete their cookies. In some businesses, reaching only a small fraction of interested buyers is still very useful; in others – especially where there are relatively few buyers to begin with – the marketer may be forced to run her own outreach programs to capture as many as possible.  In that case, the intent vendor's list would probably not include enough additional new names to be worth buying. 

How do you use it? More ways than you might think. It’s tempting just treat intent-based lists as sales leads.  But often the quality isn't high enough for that.  So the intent lists are often considered prospects to be touched through email, targeted advertising, and other low-cost media. Similarly, retargeting ads can be used to make hard sales offers or to more gently present brand messages and name-capturing content. Other applications include using presence on an intent list as a data point in a lead score, reaching out to dormant leads or current customers who suddenly register on intent lists, and tailoring messages based on the which topics the intent vendor finds an individual is consuming.

How do you test it? On the simplest level, you just apply the intent data to whatever type of program you’re testing (sales qualification, prospecting, lead scoring, reactivation, personalization, etc.) and read the results. Where things get a little tricky is figuring out which of the names would have registered as leads through some other channel, since they should be excluded from the analysis.  Similarly, you need to carefully test how new programs like lead scoring, reactivation, or topic selection programs would have performed without the intent data – it may be the good idea was the program itself. As a general rule of thumb, expect your own data to gain power as you build a longer history, so intent data is most likely to prove valuable on names early in the buying process.

Who sells it? Consumer marketers have a wide variety of intent sources, including Nielsen eXelate, Oracle Datalogix, and Neustar for lists and AdRoll, Retargeter, Fetchback, Chango and Magnetic for retargeting. B2B marketers have can work with Bombora, The Big Willow, TechTarget, IDG, and DemandBase.

Republished with author's permission from original post.

3 COMMENTS

  1. Hi David

    A very insightful post about ‘intent’ data. The problem is that many of the data sources you highlight, may not be very good indicators of actual customer intent at all. And used inappropriately, as they often have been, are starting to have unintended and undesirable consequences.

    As Nicholas Carr suggests in an article in The Atlantic on ‘Is Google Making Us Stupid?’, we increasingly use the internet as a real-time memory prompt rather than remembering things themselves or carefully filling them away. I know that I do. If I use Google to search for a roof repairers – which I did recently because it was quicker than looking through my Rolodex off business cards for a phone number – it is clearly not a guarantee that I am in the market for roof repairs. Similarly, if I visit your website to look at roof repairs it is also no guarantee that I intend to ask you to fix my roof. Obviously, if I make an enquiry about repairs, that is a very different thing. And therein lies the heart of the problem with intent data. Much of the data is about loosely implied customer intent (I think the customer is in the market for roof repairs) rather than direct customer intent (the customer has just enquired about getting a roof repaired).

    As the old saying goes, ‘all models are simplifications (of the real world), but some are useful’. Loosely implied intent data is widely used in the programmatic marketing you mention. And it works. But it is not without unintended and undesirable consequences. Most customers recognise that the data collected about them by Google, websites and social media websites is part of an implicit trade-of for a more customized experience, more customised content and other free services. Why would they give-up their data otherwise? However, as research by Turow et al on ‘The Tradeoff Fallacy’ shows, customers increasingly recognise that they are not getting hardly any of the benefits they were expecting in exchange for their data. As research by Aimia on ‘Irrelevant Marketing from Brands Gives Rise to the ‘Deletist Consumer’ shows, the failure of marketers to deliver their side of the bargain in increasingly driving customers to go ‘dark’ on marketers, by opting out of marketing, by blocking ads and by anonymysing their search.

    Loosely implied intent data will never be as useful as direct intent data. But it is all that most companies have to go on. But marketers have to recognise that customers expect something in return for their data. And customers have shown that they will go dark if it isn’t forthcoming.

    Graham Hill
    @grahamhill

    Further Reading:

    Nicholas Carr
    The Atlantic
    ‘Is Google Making Us Stupid?’
    http://www.theatlantic.com/magazine/archive/2008/07/is-google-making-us-stupid/306868/

    Turow et al
    University of Pennsylvania
    The Tradeoff Fallacy’
    https://www.asc.upenn.edu/sites/default/files/TradeoffFallacy_1.pdf

    Aimia
    ‘Irrelevant Marketing from Brands Gives Rise to the ‘Deletist Consumer’
    http://aimia.com/content/aimiawebsite/global/en/media-center/news-releases/viewer.html/en/irrelevant-marketingfrombrandsgivesrisetothedeletistconsumer

  2. Thanks Graham. I wholly agree. Clearly marketers need to test each source of intent data for each application to find what will actually work. But you raise a more interesting issue — should they use sources that do “work”, in the sense of yielding profitable promotions, even though those promotions will also reach many individuals who are not in-market and therefore potentially annoyed. There’s a “tragedy of the commons” element to this, in that marketers are doing things that help their own firm even if they harm the industry as a whole. That’s always a tough problem to fix; if memory serves, the only real solution is regulation, which none of us like. A less invasive approach is the one that has always applied to data-driven personalization, which is to avoid “creepiness” by not telling the customer what you know (or think you know) about them. So, in your roofing example, a marketer can offer roofing-related information without explicitly saying “hey, sorry about that hole in your roof”.

  3. Real-world translation and conversion are the keys here. As President Reagan said in dealing with the Soviets, “trust but verify”, taken from a Russian proverb poem: Доверяй, но проверяй (doveryai, no proveryai). The sources, stability, and veracity, of anecdotal and measurable intent data, are so varied, reliability and validity are constant issues. And, I’d suggest, this level of caution should apply to any stage of the customer life cycle, not just early stages. Perhaps it would be better to use John Kerry’s update of Reagan’s phrase: “Verify but verify”.

ADD YOUR COMMENT

Please use comments to add value to the discussion. Maximum one link to an educational blog post or article. We will NOT PUBLISH brief comments like "good post," comments that mainly promote links, or comments with links to companies, products, or services.

Please enter your comment!
Please enter your name here