It’s the people using it that will make or break any search application


Too often we don’t find what we are looking for when using the organization’s own search application. Whether it’s on the corporate site or the intranet. As a result, we waste time, scrolling for the content or fleeing back to Google, because we accept a somewhat saddening fact; that Google has a seemingly better overview of our content than anyone in our own organization.  

I spoke to Ravi Mynampaty, principal software engineer at Harward Business School about the challenges - and solutions for internal search applications. What surprised me most in the conversation, was Ravi’s point about the effectiveness of search engines having more to do with the people and culture in the organization than with back-end structure or even interface design.

Why can’t our search application be like Google?

Most companies recognize that quickly locating information is business critical. Yet, in a survey conducted by IntranetFocus nearly half of the participants said that they were dissatisfied or even very dissatisfied with their organization’s search application.

This is actually not that surprising since many organizations still don’t prioritize search. Or, they simply don’t understand how to prioritize it:

“What people tend to forget is that Google isn’t just one person or a team of people writing code. It’s several departments working together. Not just between themselves, but also with users and content creators around the world.”

Because the vast majority of search applications fails isn’t due to some technical backend-error but has all to do with people and content:

“Building a technically good search application requires skill and a good process. However, 80 % of search engines that fail, do so because of content. Specifically, the habits of people when creating and organizing content.”

In our search application, we trust?

Building a search application that “works” from a technical perspective is quite manageable. Building one that is effective is far more difficult. Ravi sees this time and time again, and to succeed, the people using the search application has to be the at the center of the process:

“Before we implement, we need to thoroughly understand the users. In what type of situations they are using the search application, and what they are looking for. Only then can we create meaningful criteria and group levels to ensure relevance.”

But no matter how thorough we are in the process of implementing the search application, it will need continuous tweaking, based on the searches that people are actually performing. This brings us to the perhaps most crucial aspect of any search application; the people using it. Of course, they aren’t the only ones unsatisfied when the top results turn out to be irrelevant. The search manager is equally frustrated, but there isn’t much he or she can do to fix the most common problem with search applications, which is one of culture and collaboration rather than of code.

On the technical side of things, this relates to meta-tagging. But getting people to meta-tag and keep the internal search application in mind when creating and structuring content on the website or the intranet, is about culture. It requires establishing an organization-wide understanding of how the search application works, and what is required from content creators if it is to stay effective. But if no one trusts the search application, to begin with, this will never happen:

“It doesn’t help that the search application is technically sound if no one has confidence in it. The Google logo gives people a sense of trust - which is kind of ironic when we think of how Google uses our data. Regardless, people are using Google, because they know they will most likely get the desired result. The challenge for the organization is to build that same trust in their own search application.”

So, while the technical aspects of the search application are created by a software engineer, it’s effectiveness is created together by everyone in the organization.