As an Enterprise Search expert, I get a lot of questions about Search and Information Architecture (IA). And what I've discovered is that people have plenty of misconceptions.
I'll share some of them in this article — and hopefully help you avoid some common mistakes about planning and implementing Enterprise Search.
1. We don’t need IA because search will do everything for us.
Many people think about search as a silver bullet. They think it magically makes everything findable -- and if this does not happen, they blame the search function. I've seen many companies switching between various search engines, dissatisfied with all of them. Instead of taking the time to plan and implement properly, they blame the technology – and decide to change it. But usually, this doesn’t bring success, but simply accelerates a vicious circle.
2. IA and search are one-time projects.
“So, how long does it take to have a solid good search?” I get that question often. And when I answer: “A lifetime,” I always get curious looks. Imagine search (and IA, too) as a garden. Digging, planting and watering is just the beginning. You have to keep working there: more watering, cut the trees and brushes, mow the lawn, remove the dead branches and leaves, etc. Without these repetitive tasks in every season, your garden will become a jungle very soon. Same with search and IA. If you don’t do these repetitive clean-up jobs, you’ll have an information jungle very soon. Take care of it -- and you’ll see how it grows and evolves.
3. One size fits all.
The next misconception is that you can use the very same solution as your friend’s company. But this is not true. Your data is different, your business culture is different and so are your goals. Be aware of this and take advantage of it.
4. Search is easy.
Everyone uses several search applications on the Internet, such as Google, Bing, Yelp or Amazon. As most of them are easy to use, people consider search to be easy in the enterprise, too. Actually, search itself might be easy. The real challenge is to provide good findability: to return all the relevant content without any "trash." The goal is to improve the users' performance by targeted search experience and help them with their everyday jobs, without overloading them with the content provided with search.
5. Search can be a security leak.
Search results are always security trimmed -- this means, if one has access to the content in the source system, he/she can see this in the result set. But if he/she has no access to a piece of content, it won’t appear as a result at all. This sounds reasonable. But if the permissions in the source system are not set properly, search can surface these problems. Imagine a company with an enormous file system: millions of documents organized into deep folder structures. Including these files into search is (technically) easy. The customer is happy with the outcome ... until a few days pass and the customer calls to complain, “We can find everything now.”
What’s wrong with that? Plenty. In this huge file share, there might be some sensitive files. For example, an Excel sheet with the name Management_Salary.xlsx -- containing all the salaries and cafeteria provided for the C-level managers. Before implementing search, nobody knew this file existed. Therefore, people didn’t open it (except the ones who had to work with).
But after search was implemented, as soon as someone searches for his/her manager’s name, this Excel file pops up in the results. Is this a security leak? No. The security problem is on the content source (in this case: file share). Search only surfaces it.
Information Architecture: Search's Secret Sauce
The misconceptions above are very common, but it’s easy to move past them if you recognize the benefits of a good Information Architecture. The key is to understand: IA and search are business processes, rather than one-time IT projects. They’re like gardening: It’s up to you if you want a nice and tidy garden -- or an overgrown jungle.