predictive marketing examples

Beyond intuition – Five times data conquered instinct

Understand the true value of a data driven, Deep AI-powered CRM targeting solution, like Tinyclues, with these real-life predictive marketing examples.

Data Operations Director

Time to read4 min read
January 18, 2017

So, you’ve read all there is to read about data, predictive marketing and artificial intelligence. And you’re full to the brim with theoretical explanations. Now, you’re ready for some real-life examples and anecdotes which show how data can make a difference in your marketing agenda. Here are five business examples that may surprise you.

1. The ideal distance between two holiday destinations is 250 km

This is a specific example of how data can help interpret consumer patterns. An online travel agent was looking to increase their repeat business by recommending destinations to customers who had already bought two or three trips. But how could they tell which destinations their customers would be interested in? The agent’s gut feeling was that customers who had already bought one trip will choose their next trip in the same area. Using the travel agent’s database, Tinyclues selected the customers who had already travelled 200 km from their homes and suggested three new test destinations: one 200 km away, one 400 km away and one at a random distance. The outcome: destinations at an average of 250 km away came out on top. The data confirmed the agent’s hunch, meaning that they could confidently offer their customers the deals best suited to them.

2. Cosmetics aren’t just for high earners

Marketers must question their own beliefs because they tend to project them onto their view of who their customer is. The following is an example of how data proved that a marketer’s hunch was wrong. An independent retailer was convinced that its clientele were mainly high earners. However, when the distributor analysed the database using the targeting functionalities within Tinyclues’ solution, they soon realised that high earners did not, in fact, make up the bulk of the retailer’s customers. Data alone helped the marketing teams understand where they were going wrong and then led them to realign their communications strategy.

3. Just because people play Call of Duty doesn’t mean they also play Super Mario

Better targeting means sending the right messages to the right people – every time. That is what predictive marketing is all about. Let’s look at an example of a retailer who previously sent a video-game newsletter to a huge base with disappointing results. The newsletter went out to all types of gamers, some casual (those who occasionally played Super Mario and Pokémon) and others hardcore (Call of Duty addicts). The outcome was that the newsletter wasn’t hitting the mark and the one-size-fits-all strategy wasn’t bringing home the bacon. But Tinyclues came to the rescue. By sending three different newsletters to smaller, highly targeted prospects (200,000 contacts as opposed to 1 million), the click-through rate went through the roof. Tinyclues’ solution even helped to earmark which customer profile would be the perfect match when each new game was released by analysing who had pre-ordered and what they’d bought in the past.

4. Holidaymakers who set off in February each year chop and change destination

This may seem a trivial point, but realistically how can travel agents predict how trends are going to change year on year? Using sales data from the previous year is a good place to start, but the stats show that they could do better. If you went to the Seychelles last February, you won’t necessarily be looking for the same deal this year. An agent would undoubtedly be wiser to suggest a different destination for this February.

To help an agent adapt its offers in real time, Tinyclues identified which client profile type matched each deal. Based on look-alike models (targets with the same characteristics) within a database, the solution identified new prospects in addition to those who had already bought this trip in the past. By identifying the right prospects, Tinyclues’ solution really helped to increase the volume of each campaign. In this instance, the travel agent was able to find many more relevant customers for this communication, increasing the list size from 15-30,000 to 150,000 relevant recipients, whose interest was clear from the campaigns’ ROI.

5. Adults buy Star Wars Lego, but not necessarily for their children

This example comes from an e-retailer who sells toys, among other things, and who used Tinyclues’ solution to unlock a surprising customer segment for his Star Wars Lego products. In the past, this range took centre stage in the e-retailer’s toy newsletters that targeted parents, with the preconceived idea that they would buy the product for their children. But data analysis uncovered a new, unexpected target: adults were buying these toys for themselves! Now the site can identify two different (but related) categories and send different messages to each of them. This discovery led to a sharp upturn in sales.

These five real life  examples look beyond a marketer’s intuition and show that AI algorithms can factor in immensely complicated databases that might have millions of members. And they can prove or disprove particular ideas to help fine-tune a company’s CRM strategy. Tinyclues’ algorithms are tailored to suit every sector, industry, product type and buying behaviour. So if you’re ready for a nice surprise, get in touch!

You might also like