We all know that in the near future electric cars will be ubiquitous. Even though they made up less than 0.4% of all cars sold in 2014, it is easy to see the potential. Many companies sell electric cars and the promise of an electric future, but not all cars are equal. The Fiat 500e, Nissan Leaf, BMW i3…all are pure electric cars but the fact that these only have around 90 miles of range can create some anxiety for drivers. So when you consider the Tesla Model S with 200 miles of range, you’re looking at an electric car that can have more widespread usage. Selling shiny objects is one thing, but packaging and operationalizing them, backed by a solid product vision, is something completely different.
The time to separate fact from fiction is now. The B2B buyer, or more accurately the buying center, exhibits a complex path to purchase and intent data is no elixir that gives the marketer superhuman abilities (it does sound pretty awesome though).
If you’re evaluating predictive marketing vendors then you will probably find that last statement rather trite. Well, this is what I’m going to address in this post. It may seem sufficient to have intent data, or behavioural data. But you only start to see real value once you are able to operationalize it in your marketing flows and processes. The devil is in the details, so let us unravel it.
Intent Data Behind A Predictive Model Needs To Be Transparent
Atul Kumar’s blog post titled Demystifying The Intent Behind “Intent Data” made the case that Intent data on its own isn’t very useful, but combining it with fit and behavioural data can provide a complete and powerful picture of the engagement with your targeted buyers.
Mintigo acquires raw web visitation data from several online publications and classifies this data into segments and conversation topics. Most of the intent data in the market is just this – a relevancy of content to a conversation topic. Mintigo uses machine learning to normalize this data to account for seasonality, relevancy and volatility. As shown in the image above, we are completely transparent about how these intent attributes, which we call Marketing Indicators (MIs), contribute to the predictive model. Instead of providing just a dumb “predictive” score, it’s important to understand the insights behind that score based on the data. Here’s an example of the output of what a typical predictive model, or what we call a market, would look like:
Fit & Intent Data Can Be Used To Drive Intelligent Marketing Programs
Let’s assume ACME software has launched ACME CX, a customer success software. By building a predictive model, ACME is able to identify the DNA (the output from a Mintigo predictive model) of the best customers for the CX product. An interesting attribute for the marketer is Net Promoter Score (NPS) because it has a high conversion lift as identified by the predictive model. So when visitors are reading about and engaging with content about NPS, we know they’re a good fit for the ACME CX product. So if you can export this piece of intent data into your marketing automation platform, you can perform intelligent routing in your marketing workflows. Below is an example of this workflow in an Eloqua canvas (this can also be done in Marketo as well).
Intent Data Can Dynamically Change Throughout the B2B Buyer Lifecycle
Intent isn’t descriptive for a fixed point of time. As the buyer navigates through the complex B2B buyer journey, their intent and the intensity of the intent changes. What this means is that a buyer will consume different types of content and engage in varying capacities depending on the part of the funnel they are in – discovery, evaluation, consideration. And generally the DNA associated with the different stages change. Marketers have to keep this in mind when they build predictive models using intent data – they need to ensure that they score the leads through the predictive model frequently as the lead moves through the entire lifecycle in order to build a precise and effective conversation with the buyer.
Intent data is valuable, but high-quality-modeled intent data is a lot more valuable. To figure out if you have access to high quality modeled intent data, you’ll need to ask questions such as:
- Do you know how the visitation data is being transformed into intent?
- Is the intent data being used in the model? Do you understand the results? Can you act on it by creating marketing strategy and content based on it?
- Are you able to target on the intent indicators in your marketing automation software?
- How often can you refresh the predictive models to take advantage of the latest intent?