Automated Recommendations (Finally) Hit Their Stride
Gone are the days when online businesses will throw technology at business problems. Don’t get me wrong — technology’s role in business has never been stronger, but today there’s a difference: the ‘business’ is rapidly taking over how technology is used, and is demanding tools that are friendly to marketers, merchandisers, and other biz types with line-of-business (LOB)/revenue responsibilities.
The other difference is that business folks are looking for tools that help them address specific key performance indicators (KPIs). It’s hard to find new budget dollars for technology, but it’s easy to justify investing in, for example, a tool that will easily automate cross-selling and therefore increase AOV. There’s no better example of this than personalization tools such as recommendations engines. The conversation we have with customers regarding recommendations today is about what they’re specifically trying to do at each stage of the conversion funnel, rather than one that sounds like “this year we need to add a recommendations engine to our site”.
This means that tools that automate the recommendation of relevant products and content (e.g. help or news articles) must offer a really high level of marketer control. From brand look & feel, to how recommendations are messaged, to practical considerations like the price threshold of recommended products — if marketers can’t incorporate their own best practices, the tool is unlikely to accomplish their specific goals.
Another important consideration: every square inch of online real estate counts. If a merchandiser is going to use some of the that space to show off popular articles, similar products or try to entice an impulse buy at the end of a shopping spree, they’d better be darn sure these recommendations are performing. Therefore, it’s essential that business users are able to easily try out different recommendation strategies with different audiences, test & compare them easily and make changes (and push them live — without IT).
As a testament to the growing popularity of automated recommendations solutions, we’re starting to see (overdue) attention to what I call the “new breed” recommendations products by the analyst community.
Susan Aldrich, SVP at the Patricia Seybold Group, recently evaluated Omniture Recommendations. The report provides a thorough review of key areas from her recommendations evaluation framework such as recommendations structure, managing recommendations and integration with other online marketing/merchandising components. I encourage you to download your own free copy of this report.