Improving Findability Is Not Just About Metadata
Coming back from SharePoint Saturday DC this past weekend, I was reflecting on some interesting conversations from the event around the state of search in SharePoint, especially with the pending UX advances surrounding Oslo and Office Graph, which we should begin to see trickling into Office 365 later this year. Feedback from SPSDC participants was enlightening, though: generally speaking, people are not happy with search results within their enterprise collaboration platforms (not so much about the technology, but more to do with their implementation and UX).
Within each conversation, I brought up the difference between "searchability" and "findability," with the former focusing on the interface through which we search for content (best bets, high-level site topology, keywords presented on the search page) and the latter focusing on improving the assignment of metadata, ensuring the user experience around making data "findable" is positive and, ultimately, effective at driving/improving improved search results.
I find the topic fascinating, and still believe search to be the most important aspect of any knowledge management platform. If you can’t find your content, what good is the platform?
Daniel Tunkelang (@dtunkelang), head of query understanding at LinkedIn, discussed the frustration of end users in a presentation last summer (2013) entitled Enterprise Search: How do We Get There from Here? and is definitely worth a read as you start to outline your own path forward to move from a focus on Searching, and instead focus on Finding. (I especially like slide 9 of his deck, which encapsulates the reality of the metadata problem in most organizations)
What is Searchability versus Findability? Is it just semantics at play, or is there really a difference between building out a solution focused on finding the right content rather than just searching for content? Is it wordplay? Somewhat, but there is real meaning behind it — and it is something all organizations serious about improving the search experience for their customers and employees should consider.
Most organizations approach the problem by gathering all content into one location and slapping a search tool on top, assuming that this will provide a stellar search experience — or at least solve the problem quickly so that they can move on to other issues. Unfortunately, this is rarely the right approach.
There are multiple problems with search. Few organizations get it right. It’s what my good friend and consultant Paul Culmsee (@PaulCulmsee) calls a "wicked problem" in which there may be many paths forward, but the longer you delay taking action, the more complex and ugly the problem becomes. Rarely does ignoring a problem help solve that problem — and search problems are quickly compounded by fast-growing content databases and an ever-expanding corporate focus on collaboration. The volume of content in your organization is likely not decreasing, but rapidly growing. Building a search entry point into that content is a company imperative.
A speaker at a SharePoint event once described the Searching versus Finding problem as "If your child is lost, do you want to search for your child, or find your child?" Obviously, the goal of building out your search capability is not about simply having search, but in helping people find the data they need quickly and in a format they can use.
The key to developing a strategy focused on Findability is more than just expanding the use of metadata, but in understanding the end user search experience end-to-end:
- How do people search?
- What is the expected result?
- Was the search successful?
- How do you capture feedback?
- What do you expect end users to do to add to the process to help improve future searches? (metadata)
- How much should you automate?
As you think about your own SharePoint deployment, and talk to your end users about their search expectations, familiarize yourself with the search features in SharePoint 2013 on TechNet, as well as specific guidance on planning for your search implementation.
The number one complaint about knowledge management platforms from end users is that people are unable to find their content. The most successful SharePoint deployments are the ones where the time was taken to understand end user expectations, and build to meet those expectations. Post-deployment, even your ongoing governance activities should be centered around end user feedback and priorities, because at the end of the day — if people can’t find their content, they’ll go elsewhere. Search is a key ingredient to driving value out of your SharePoint deployment.
Nice post, and it states what I have been working with for the past 10+ years. I also enjoy Daniel Tunkelangs work immensely.
Seems in the Enterprise the challenge is the same today as when I started out with search, and I use the same techniques today as well to make it better. How do people work, how can search as a technology help you do parts of your job better? I never ever talk about search in a search project unless someone else brings it up. I want to know how you get the information you need to solve tasks. Ask a colleague, use tarot cards, or perhaps search? 😉
As for SharePoint the tools has gotten a lot better, especially with 2013. Today it is so much quicker to create context aware search and pretty good UI’s, without doing a ton of development – which makes it easier to do A/B testing and solve the problems instead of doing plumbing.
Indeed, and even just a little extra in the way of social behaviors goes a long ways, as Google’s search competitors in the late 1990s/early 2000s found out. The simple notion of an algorithm leveraging web page linking behaviors to augment classic text matching provided the extra quality that put all of Google’s competitors out of business and led to a few hundred billion dollars of market value.
It’s also why social has become so important to search. There’s only so much you can anticipate and automate. An algorithm can never (not yet, anyway) match the human brain’s ability to connect (or create) patterns in content, but by including social interaction within the mix, you are more likely to surface relevant content.
I think that’s why Microsoft is putting so much emphasis around Office Graph and Oslo as the future of their search interface — its about presenting you with the most relevant, personalized data possible. I don’t think it totally moves us away from some kind of search interface, but there’s a ton of room for improvement.
Agreed. If people can’t find their content, they begin taking content storage into their own hands, building new information silos and perpetuating many of the problems in knowledge management.
Of course, I’ve got a little bias, but I kind of like Sympmarc’s sentiments, “the best search is the one you don’t need to do,” and which is echoed on slide 29 of Daniel Tunkelang’s presentation, “Design an experience that doesn’t require search.” In other words, strive to anticipate what a user needs when at all possible (but recognize search will always be an important part of the overall findability solution).
That ability to flexibly anticipate is, of course, now becoming easier because we can in a sense flip search around–rather than just people searching for knowledge, knowledge can also be continuously “searching” for the users who most need the knowledge at a given time. And to continue the analogy, the “meta-data” associated with people that the knowledge searches against are each person’s particular inferred interests, expertise, and experiences.
I’ve always believed that the best search is the one you don’t need to do. Findability (which isn’t a word!) is about creating multiple ways to get to the right content in the context of doing the work. That’s why things like the Content Search Web Part are so important in SharePoint 2013. Embedding content in the right places – whether it is search-driven or not – leads to more streamlined knowledge management. More importantly, it can lead to improved performance, which is the actual goal.
M.
Excellent post. Thank you. My only comment is that in the enterprise if people can’t find their content there is no where else to go. They end up either interrupting someone else’s work or making a decision based on inadequate information. Search is business (and career) critical