If you’re a B2B marketer, I bet you’ve heard of the term “intent data” recently. There’s a heck of a lot of intent in selling “intent data” right now. Pitches such as “we can tell you (because we have intent data) who will buy, what will they buy, when they will buy and what will they pay” are music to any salesperson’s or marketer’s ears. Hey, if this is entirely true then why do you need marketing or sales? Just get bunch of order-takers and bingo — happy CFO, happy CEO and hopefully happy shareholders.
The truth is that it takes a lot more to make a prediction of who will buy, what they will buy and when they will buy. Let’s start with the following scenario:
A company needs to provide real-time analytics to enable their business. The CFO approves the budget for the next fiscal year and the project gets started. The IT team begins evaluating whether to build or to buy. They research several database options as well as many of the analytics vendor offerings. The process can take anywhere from a few weeks to several months to finally decide and perhaps make a purchase.
What is evident from our example is that a typical B2B buying process starts with the change in status quo (i.e. a need for real-time analytics). It is then followed by a commitment (someone with budget approves the project). It is true that sometimes a project can start without budget and the budget requests are then made later in the process (hint – a lot of tire kicking). The rest of the buyer journey includes extensive research (reading stuff online, offline, asking friends, industry experts, analysts etc.), exploring options, committing to solution(s), evaluating solution and finally making the decision. The process looks something like this:
Of course there are variations in this process, especially the time it takes to complete purchase. It really depends on the type of the product, urgency of the need and the size of the buyer. For example, Sony Pictures bought security services from FireEye within a month of getting hacked. Similar purchases without a compelling event (such as being hacked!) can take several months or quarters. The need and the compelling event can influence the purchase cycles dramatically.
The initial phase in the B2B buyer journey is the ‘Discovery’ phase. The stakeholders do tons of research mostly online. In our example, the IT team members will visit several industry publications, read about databases and explore pre-packaged software options. Such events (over a billion a month in volume and growing) are captured by publishers and ad-tech vendors such as Bombora, Magnetic, etc. Over 99% of web visitation events are anonymous and only 30% to 40% are resolved at the company level. The typical data set includes the following elements and others, such as time and location information:
|Topic of the web page visited||Database||Real-time analytics|
|Relevance of the topic||0.21||0.04|
What this data by itself will indicate, at best, is an interest. This interest can be identified because someone likes to read about databases and real-time analytics, or it could be because there is a need…or both. The simple act of someone from a company reading last week about a particular topic (i.e., web visitation events) does not imply a purchase intent in the B2B buying process. Over 400 million (and rapidly growing) monthly web visitations contain a lot of noise — noise being interest or random searches and not necessarily the intent to buy. The point is that you don’t want to make your marketing decisions based on just web visitation data alone…which many are erroneously calling intent data.
To contrast this, the B2C buying process is rather simple. You have a need or desire, you do some quick research, and then you buy. As a real-life example, our seven year old dishwasher stopped working last week; I did a bunch of research, shortlisted our choices, my wife verified and approved these choices, we found the best price, and then we made the purchase. All within 72 hours. So if you were to examine my web visitation history, the noise for dishwasher (a topic) is nearly zero (admittingly, I did my research during the 4th of July anticipating the imminent death of our dishwasher). As a result, my recent online research (i.e., web visitation data) was a huge signal for our intent to buy a dishwasher in the near term. However, this is simply not the case in B2B buying cycles. As discussed above, it is usually a lengthy process (few weeks to several months) and it takes a team of stakeholders to make the buying decision — thus what is viewed as intent data may contain a lot of noise.
True value surfaces when you use scientific methods to detect purchase intent from raw web visitations data. It’s a signal to noise problem. A commonly-used approach is to use the ‘share of voice’ concept — web visitation data constantly analyzed over time and evaluated for visiting companies to determine their purchase intent. This is indeed an oversimplification (and perhaps somewhat inaccurate) of actual machine learning and mathematical models used by Mintigo to determine purchase intent.
Now that we understand what purchase intent really means, the next question to ask is this: is knowing the intent alone sufficient to predict who will buy from you? The answer is obviously ‘no’. B2B vendors have a very specific customer profile for a given product and geography. Only certain kinds of customers buy from you. Understanding the real characteristics, or the “DNA”, of your customers is critical for marketing and selling. The DNA of your customers acts like a matchmaker.
For example, Mintigo’s customers are early adopters of technology and are modern digital marketers. So there are companies that may have an intent to purchase predictive marketing solutions but are not in our sweet spot — meaning they don’t match our ideal customer DNA. Finding the right fit prospects (i.e., those with the right CustomerDNATM) with demonstrable buying intent is the key to success. Otherwise you can spend your precious resources in chasing down prospects that will never buy from you.
The last thing you need to understand is who is already engaged with you — the level of engagement can have an impact on a final purchase decision. Many marketing platforms such as Oracle Marketing Cloud, Marketo, and Adobe Marketing Cloud include ways to measure the level of engagement via activity-based implicit scoring.
In summary, determining real purchase intent is a complex process requiring expertise in machine learning and other statistical methods. Merely using a single data point will lead to bad decisions. You need to use the combination of fit (CustomerDNA), intent, and behavioral engagement data to find your golden nuggets.