Microsoft introduced Clippit—a character known infamously among its users as “Clippy”—in 1996, not long before the 100th birthday of the paper clip. This intelligent-user interface was the best-known iteration of Microsoft Office Assistant. Clippy hung in there until 2002 and finally vanished from the Office suite for good in 2007, but it remains immortalized in memes. Even Clippy’s designer admits to creating a character that “still annoys millions of people every day.”
Clippy’s failure suggests a misunderstanding of user needs. Instead of enhancing productivity, Clippy interrupted it, and gained a reputation as unhelpful, unwelcome and, to many users, downright unpleasant.
Clippy had a worthy goal: to increase engagement with Microsoft Office users while they were using the product. So why did it fail so spectacularly?
Clippy didn’t have context. It didn’t fully understand user behavior and modify its interaction accordingly. For example, when you typed “dear,” Clippy wanted to help you write a letter. It just didn’t understand that you may not have wanted to write a letter in the first place.
Since then, data collection and analysis have made it possible to better understand end-user behavior and activity so that communication within a product can be optimized to drive and enhance user engagement. Compare Clippy with a more evolved form of data-driven in-application messaging, Amazon’s recommendation engine. Amazon’s algorithms pull together massive amounts of information on searches, purchases and more. And because the information Amazon shares is often helpful, users continue to engage with Amazon and its recommendations. This results in more purchases, more data collection and a greater ability for Amazon to further fine-tune its recommendations.
Today, software product managers can gain this insight and provide context through a powerful combination of software-usage analytics and in-application messaging. Successful user engagement is possible by driving dynamic messages that are contextually relevant and rooted in learnings from usage intelligence. By collecting data on feature and product usage, hardware and OS metrics and more, usage analytics provide the insight needed to set the context for in-application messaging: who to communicate with, when to reach out, what to say, and how to listen and process information….