We recently met a well-known marketing agency.
When they learnt more about what we do, they revealed that they also use intent data, from another provider in the market. This was exciting news! We love geeking out about intent data and its many use cases. But then they said something that surprised us:
‘Isn’t intent data just an educated guess? We find it can be so hit or miss.’
For a moment, there was a very heavy silence that filled the zoom call.
In all honesty, we were taken aback to hear all intent data being painted with the same brush. After all, all intent data is not created equal.
The idea of intent data as a guess that’s hit or miss just didn’t sit well with us. Certainly not our intent data – not only do we use it ourselves, but our customers continually get amazing results. And that’s why we’re writing this post – to share our approach to intent data and the work that we put into our platform to make it reliably useful, accurate and relevant.
Working to create the best intent data
Let’s flashback to a few months ago. Covid-19 restrictions had been lifted and the Cyance team was meeting in person for the first time in ages. Our CPO and Founder, Jon Clarke, and our Product and Support Director, Claire Gardner, were going to deliver a presentation on the new iteration of Cyance we were due to launch soon. Suffice to say, there was an air of excitement to the meeting.
One of the key points that Jon and Claire emphasised over the course of their presentation was the new algorithm our data scientists had been working on. It had been developed to provide users with a picture of buyer intent over a period of time, which is a more accurate reflection of intent to purchase than one-off surges in intent.
The new algorithm would not only improve the user’s understanding of the problems their prospects are trying to solve with products and services like theirs, but would also improve the precision with which we can identify good-fit matches with our Ideal Customer Profile criteria, where the prospect is in the buyer journey, and what content or messaging would resonate with them most.
However, Jon, Claire and our data scientists still weren’t 100% satisfied with an element of the algorithm and needed yet more feedback from our Customer Success & Sales teams – who are all experienced marketers that live and breathe intent data.
When the platform states a company is exhibiting intent to purchase, the algorithm takes into account buyer interest or activity that has been consistently shown across a period of time. But what patterns and ‘durations of intent’ balance the need to discover genuine early-stage buyers, while filtering out false-positives of companies who aren’t serious about finding solutions to genuine problems?
We were so keen on launching the new product as we felt it already had much to offer, but we didn’t want to compromise on our company value of delivering only the best intent data for our users – not the ‘noisy’ or misleading intent data that so many others providers dole out. (If you’re unsure about the accuracy of your existing intent data, then perhaps consider asking your intent data provider key questions.)
Another issue that the Product Team was working on was the amount of intent data showing in our platform over the last few months.
As you may be aware, Cyance uses thousands of online sources to track and measure buyer behavior. Many websites and publishers from which we received intent data had paused sending signals until they resolved their consent management issues in line with changing privacy regulations, and rightly so.
As a consequence, many of our competitors simply lowered the threshold for what is considered an intent signal, in order to maintain the level of data in their platforms, something Cyance never considered. We took the strategic decision early on never to compromise on the quality of our data (for example, if we consider 120 seconds time spent on a relevant article by a user in a company as a data point for our algorithm, we wouldn’t reduce it to 60 seconds just so that the platform displays data to our customers).
However, we needed to improve the quantity of data in our platform while maintaining our focus on the accuracy and quality of our intent signals.
So we went back to the drawing board.
An algorithm that maximised the benefits of intent data
We worked hard to ensure the algorithm now surfaces two types of research behaviour trends – the ‘bump’ and ‘crown’ cycles, to distinguish genuine committee-based buyer behaviour from false positives.
The ‘bump’ research cycle is indicative of research conducted by individuals tasked with finding solutions to a business problem or goal within a typical 12-week cycle – after which they present their findings to key decision-makers. The ‘crown’ research cycle is indicative of sporadic research conducted by members of a buying group across a 12-week cycle.
This sporadic approach is often due to the members needing time to focus on other tasks as well. Though intent activity may vary between the ‘bump’ and ‘crown’ research cycles, the critical time frame is typically 12 weeks. Our algorithm was refined to factor this in.
This is just one example of many research-based and data-driven refinements made to both the algorithm and our platform to accurately reflect buyer intent. We drew upon data science, trend analysis and anonymised aggregated data. Then we ran tests. A lot of tests! We wanted to make sure we got it right, so we were meticulous in our approach.
Not only that, we spoke with our users and fine-tuned our platform to give them the key features they want:
- To find companies that are a good fit and research business topics relevant to the user, within a specific date range
- To find the most popular business topics relevant to the user, and the companies researching these topics, by date range
- To track companies that are a good fit and the topics they are researching, within a specific date range
Focussing on Quality and Quantity
In order to raise the level of high quality and accurate signals in our platform, we took the following steps – which we are delighted to say have improved the overall quality of intent data in our platform. In fact, our data is even better than it was before the increased privacy regulations. It’s true what they say – out of adversity comes opportunity!
- We have implemented an improvement to how we match cookie events to companies. Upgrades in the matching algorithm and new company data sources also help with this crucial identification process.
- More publishers and websites are providing us with intent signals as they resolve their consent management issues in line with TCF V2.0 standard. We wouldn’t have it any other way.
- We have improved our publisher relevance scoring methodology and added yet more high-quality publishers to our database. This is helping to find more relevant lookalike publishers and also tune out bad fit events.
All these measures have resulted in an approximate 138% increase in quality data on our platform, compared to the pre-privacy regulations data.
As a result of all of the above, we are deeply honoured to announce that as of last week, one of the Big Four global accountancy firms has decided to use our intent data to fuel their sales and marketing, as well as a well-known multinational technology company.
So, is intent data worth it?
It’s impossible to ignore the results of hard work and constant refinements. Without a doubt, we can say we deliver on what we promise – the most accurate and relevant reflection of buyer intent.
And that’s why we don’t agree with the statement that intent data is an “educated guess”. Well, maybe it is if you’re relying on intent data from other providers.
But with Cyance, our customers’ results speak for themselves – and that’s not just because we’ve got intent data down to a science, but because we’ve done our utmost to make intent data into a science.
T0 learn more about our unique approach to intent data, book a demo