Imagine you want to know the time of the next train between two cities. When you type this into Google, the first hit isn’t a link to a site but a card like the one below. It not only gives the times but also useful additional information: a map, trip duration, and tabs for walking, driving, and cycling. Enterprise search isn’t like this. The same query on an intranet gives the equivalent of a link to a PDF containing the timetable for the whole region. It’s like saying “here’s the book, look it up yourself”.
This is not only a poor user experience for the employee, but a direct cost to the employer in wasted time. I’d like to see enterprise search move away from results pages of links to providing pages of answers too, and cards are a powerful way of doing this.
The appeal of cards
There are a number of reasons why cards work well as a design pattern:
- The information is self-contained, and they highlight what is most important
- They are ideal for mobile devices (think Google Now and iOS Passbook)
- They work well for responsive designs – larger displays can tile multiple cards
- They can be manipulated: metaphors such as flipping the card over to see more detail or expanding sections work well.
All of these factors mean that cards are becoming more common in consumer interface designs: Twitter, Flipboard, Facebook and Pinterest are all developing in this direction, meaning that people in the workplace will be more familiar with them too.
A practical example
Let’s take a more typical enterprise search scenario: an employee has had the happy news that he is to become a father so wants to know if he is entitled to paternity leave. In many cases, the search may not work at all because the answer lives in a separate HR system, or the content is labelled “parental leave” so there is no match for “paternity”, but that’s a whole other set of underlying problems that I won’t dwell on.
Let’s assume that the search does find a result. What is shown will often be a link to an employee handbook document – sometimes several hundred pages long. He will then have to open the document and repeat the search to find the right paragraph. A mock-up of how this looks in SharePoint is shown below, but the experience is similar on other search applications too. No wonder we end up repeating all this information as FAQs (and there are other good reasons why FAQs should be avoided).
We could improve this using ‘best bets’ by forcing the first hit to contain useful information. In the mock-up below, the keywords ‘paternity leave’ force the hit to be a very short document, but it contains far less information than the card concept we started out with.
To make the concept really work in an organisation, we need to do more than simply pre-package answers. To use the small card space effectively, the answer needs to be personalised to the individual. For example, leave policies often vary by country, so the answer will need to adapt to the employee’s employment location. It may even change according to years of service or grade.
It’s not only the employee’s profile that could be used to tailor the card though. Just as Google Now uses your location to look up ‘get me home’ details, a card could be GPS-driven. As a sales employee nears a client, the mobile-friendly intranet card could show recent orders, press coverage, and open questions from the client to the support desk.
Cards could be used for things like company products too. The first hit could be an image, sales data, key information from the CRM system about main customers, contact details for the brand manager in the employee’s region etc. Note that both these scenarios involve merging data from multiple sources; cards are not just a fancy way of showing hits.
Being realistic, enterprise search is rarely at the forefront of company investments and innovation. The first move to enable something like this needs to come from search engine and intranet vendors. Microsoft is making some moves in this direction with Delve and its Next Generation Portals. Other vendors do a good job of using cards to display employee details (e.g. Intranet Connections) or conversations (RightPoint Spark), but none have focussed on making it easy to encapsulate succinct content.
It will also require content owners to think in a new way about how their content gets delivered. The perennial complaint about enterprise search is that the results are irrelevant, but in reality it’s often because the answers haven’t been created.