Role of intent inference in business intelligence

When I met Harish Reddy a few weeks ago, it was ‘LeadForce1’ that was prominent on his card. And Harish, who heads the APAC operations from Bangalore (, spoke about his company’s work in ‘Marketing Automation 2.0’ or ‘next-generation marketing automation that converts anonymous online visits into qualified sales leads, determines website visitor interest and intent, and enables sales teams to reach decision-makers more effectively and close deals faster using patented business intelligence and data mining technology.’

Our conversation continued over the email, with the difference that his company’s name now is LeadFormix, ‘because of objection from Salesforce.’

Excerpts from the interview.

First, an overview of how, over the last few decades, business intelligence gathering has evolved.

In a 1958 article, IBM researcher Hans Peter Luhn first used the term business intelligence. He defined intelligence as: “the ability to apprehend the interrelationships of presented facts in such a way as to guide action towards a desired goal.”

In 1989 Howard Dresner (later a Gartner Group analyst) proposed BI as an umbrella term to describe “concepts and methods to improve business decision-making by using fact-based support systems.” It was not until the late 1990s that this usage became widespread.

Today, BI gathering manifests itself in many types and forms, especially in the online world. From understanding the keyword preferences of users to boost business opportunities, to using eye-tracking methodology to make advertisements more visible and effective, to social media networks using user profile data to target promotional campaigns – these are all various forms of business intelligence gathering techniques that companies use for adding more value to the services they provide.

While in the initial years of BI gathering, the focus was more on web analytics – providing ample information about past performance such as page views, click-throughs, time on site, and so forth – with time, Google added more specific elements such as the source of the visitor (the website he came from), repeat visits based on IP tracking etc. While this information was critical from a metrics perspective, it was not enough to provide businesses with insight on what to do next, or how to make things work better.

And this is when BI gathering started undergoing a monumental change; it slowly started moving towards behavioural analytics. Simply put, behavioural analytics is the ability to follow patterns and uncover similar sequence of actions by clusters of people, such as, what they look for once they hit a website, what they search for in their subsequent visits, the kind of content they download on their different visits etc. This can give a much deeper insight into understanding the psyche of the target audience of a brand.

The payoff from understanding these behavioural patterns is huge: you can use this insight to optimise your site, influence people’s behaviour, and ultimately drive actions that you care about – clicks, page views, sales, and so on.

The latest technology in behavioural analytics is finding the intent of visitors to a website. The implications of having such information on your visitors can be drastic for any business, arming them with enough data to know exactly what their customer is looking for and provide for accordingly.

Any examples of businesses using and gaining out of this process?

Intelligence gathering is critical to any business whether it is selling to customers (B2C) or to other businesses (B2B). Customer surveys, research reports are all different forms of intelligence gathering techniques adopted by businesses to understand the needs and motivations of their end customers.

In B2C it is easier to measure consumer reaction to a product, since the buying decisions are taken individually and spontaneously. You hardly hear B2B success stories where a product sold 1 million pieces in the first month, while it might be a regular scenario at the Apple headquarters.

Hence intelligence gathering and if possible inferring the intent of a business lead become a priority for a B2B company, where sales cycles can last as long as 12 to 18 months.

Intent inference has many business use cases. But here I would like to highlight two, which affect the sales and the product division in a company.

For the Chief Sales Officer:

In the absence of intent intelligence, B2B enterprises use primarily two ways to sell their solutions – they either put together a list of all the companies which can use their product and start approaching them over emails and cold calls, trying to get a breakthrough on someone who might be interested in buying their solutions. Or, they respond to inquiries they might get on their website, email or calls.

In both these scenarios the sales cycle is completely dependent on the reaction of the person, who the sales person approaches or the person who approaches the company with an inquiry. Result is that the sales cycles become longer and the rate of conversion, lower.

Instead, having intent intelligence helps companies to approach their leads proactively and drive the sales cycle. Today, having a website is integral to any business. In fact adding website address to all our marketing collateral is almost a given. But what companies do not realise is that despite driving traffic to their websites, they are not gathering this lead data and pursuing them, converting them into sale opportunities.

Gathering intelligence on these visitors not only ensures that the company’s sales funnel is full with leads who have shown some level of interest in the product, but also helps in knowing exactly what they looked for on the website, the solutions that interested them, giving the salesman enough dough on what to sell to the visiting enterprise when he approaches them, and thus resulting in shorter sales cycles and higher conversions.

Some of the Fortune500 companies which started using the intent intelligence gathering technology saw a 100 per cent increase in the volume of leads they generated from their website. Also, their sales teams found that following and converting leads which were interested in a solution similar to their company’s offerings was much easier than making several cold calls to potential customers who may not even have a requirement for the solution they sell.

On the whole, a deeper insight into visitor intent is critical to the success of any lead management process. Understanding and mapping visitor intent results in:

* Better qualification of leads.

* Enabling marketing to formulate effective lead nurturing strategies and run intent driven campaigns.

* Ensuring relevant and customised communication and messaging resulting in shorter sales cycles.

* Enabling sales to focus on sales-ready prospects only and in real time.

* Empowering sales with key information regarding what the prospect’s needs are and what solution to pitch resulting in faster conversions.

For the Chief Technology Officer:

Similarly, from a product development perspective, the data provided by inferring the intent of a visitor can help in understanding what features or solutions are a hit with the business leads. The kind of information they search for on the site coupled with the final sales data, can help the company CTO to decide on which solutions and technologies need to be scaled or upgraded and which can be shelved. Since technology is evolving at a fast pace, it is important to be in the know of what is attracting customers and what technological advancements will keep them coming.

Intent inference helps with finding these motivations of customers which they unknowingly express on the site, while browsing for information.

How are you going about gathering the intention and what are the challenges?

Inferring the intent of a person depends on the analysis of several factors such as the sequence of activity on the website, the pages and phrases interested in, actions performed on the website and time spent. Such behavioural tracking and analysis, especially when combined with historical information and when seen from a company level decision making view, results in a clear understanding of:

* What the visitor is looking for as a solution, and

* Where the visitor is in his buying cycle.

The challenge lies in ensuring that the patterns derived are statistically accurate. Since every visitor is different and the behaviour different, it often becomes difficult to infer decisions based on keywords expressed as intention. The real challenge lies in correlating the available data with a pattern and then arriving at an accurate conclusion of intent expressed. Having said that and having processed data in the range of millions of intentions, it becomes easy to decipher patterns and group that behaviour.

Can you describe the core of the system used for intent gathering?

The core of the system is an unsupervised learning system that uses probabilistic models and self-learning algorithms to derive patterns from the web behaviour CRM data that have been gathered. The data are mined and predictive models are built which help determine the intent of the visitor company to the site and to plot the current sales stage of that visitor (research vs buy). These rules are configurable and can be customised to the specific nuances of sites.

The algorithm, among other things, considers not only the behaviour of a particular visitor to the site but also extrapolates this with the behaviour of the entire company to give a holistic view. This is combined with the pages that are viewed, the importance of the page (a career page vs a data sheet download), the depth of information accessed, and the time spent. This information is combined with other marketing responses such as email, webinar, content consumption etc. to provide an aggregate intent level. This is calculated in real time and the sales stage is updated as and when the behaviour indicates this change.

On cutting edge developments in research, and what we can expect in the near future from this technology.

Cloud-based data mining and statistical analysis are just getting started and we will have many new applications of the same. The advantages of cloud-based systems are the amount of resources that can be used, and their distributed nature.

The techniques used for detecting intent are generic and hence can be applied to various other fields. In marketing itself, real time analysis of visitors can be used to customise the website for a particular type of visitor (target segments), matching the content with the viewer.

Search results can be tweaked to match the viewer – while most search engines concentrate on providing answers to the ‘What’ query, the answers to ‘Who’ adds another dimension and personalises the search more.

Pattern detection can also be used for recognising any odd behaviour and to look out for unnatural activity – converting it into a fraud detection technology.

The use cases for intent technology are many and the list of potential applications is quite generic and very widespread.


Our code of editorial values

This article is closed for comments.
Please Email the Editor

Printable version | Jan 24, 2022 9:55:46 AM |

Next Story