Connect Mintel’s 35k+ global, unique monthly users to the market intelligence reports they need to support their business initiatives, giving value to their Mintel subscription.
Using Fusion query pipelines, provide search results that speak a user’s language, and offer relevant recommendations that increase customer satisfaction and the value of a Mintel subscription.
In the initial rollout, Fusion reduced query response times by 90%. Now, with customers finding reports they need, and recommendations providing reports they didn’t know they needed, retention is high.
When serving not only the incredible breadth of subject matter in Mintel’s reports, but also the expanse of languages and locations of its customer base, providing an intuitive experience for each customer can feel overwhelming. Fusion query pipelines make it possible to quickly deliver relevant results to every user.
With customers concentrated in Thailand, Germany, China, and Brazil, Mintel offers reports written not only in English, but also in Thai, German, Mandarin, and Portuguese. Each piece of content is bundled into one document with alternately named fields like title.lang.en for English and title.lang.de for German. Using the language specified for the user interface as an argument that’s passed to Fusion, a query pipeline adjusts the fields it searches and the data it returns to the user based on his preferred language to ensure that a Portuguese speaker isn’t given a report in Mandarin.
High-value content, or any content that the marketing team specifies should be given high visibility, can be tagged as “featured.” When a user’s search terms match featured pieces, a query pipeline will boost the most relevant of those pieces to the top of the search results.
Similarly, since research findings in the verticals Mintel serves can change rapidly, publication dates are essential to producing relevant search results. However, the shelf-life of a report depends on what type of report it is, so a time decay function is applied to appropriately boost or bury content based on type. For example, one of Mintel’s in-depth reports has longer legs than a frequently-published topical insight piece so the latter is buried in search results by the time decay function sooner than the former.
By defining query pipelines in a modular way, developers can easily drop reusable steps into multiple pipelines. And testing in Fusion before changes are pushed to production, to compare how search results have moved up or down when tweaking the pipelines, reduces the development time required to get the desired results.
Once a user has clicked into a report, he’ll find recommendations for other meaningful content that could be useful to him. To produce these recommendations, Fusion algorithms rely on machine learning that takes into account the category tags of the report the user is currently reading and what they’ve read in the past, suggesting articles with categories that relate to one or both of these groups.
Content is Mintel’s bread and butter, and with relevant recommendations, both the customers and the company benefit. The customer, finding more data to back his research, gains more value from his Mintel subscription, and Mintel, whose customers are happier with the applicable data they find, has customers who are more likely to renew their subscription.
In 1997, Mintel became the first market research supplier to provide online access to its content. More than 20 years later, they know what their customers want. “Search is what drives engagement with content,” says Adrian Rogers, AVP of Engineering at Mintel, “and when people buy a subscription, they need to find the content they’re looking for. Fusion’s recommendations and query pipelines help our team provide the advantageous experience our customers require.”
Whether it’s that one piece of data that lives in the online version of what was once produced as an inch-thick book or last week’s update to current beauty trends, Fusion guides Mintel’s customers to the right answers.