Why Does Enterprise Search Have Such a Negative Reputation?

KMWorld 2019 review

This month is my six year anniversary at BA Insight, and during these years, after many discussions with prospects, customers and even analysts, it is obvious that enterprise search has a bad rap. I’m not sure if you have observed this, but convincing leadership that enterprise search should be an area of investment is often daunting. In fact, an analyst from a well-known firm recently told me that it is much easier to justify search for customer support than it is to justify it for the broad employee base because you can actually measure call deflection. He then added that the same is true of site search since everybody knows that you need your website to be up to snuff or customers will abandon you. I countered that if 50% of employees saved 5 minutes per day when looking for information, and there are 1,000 employees, then the savings equate to 2,500 minutes per day (~42 hrs.). This equates to 600,000 minutes per year (10,000 hrs.), which is a savings of $500,000 per year at $50/hour – which is about what a $100K person makes.

The counter from the analyst was that nobody measures employee productivity, and although you are right that the savings is there, and it is probably even a lot more than you suggested, we don’t know what employees will do with their time.

Isn’t it surprising that when we get to discussions like this, suddenly we don’t trust our employees? It is kind of like “let’s make their work life hard and miserable because if we helped make them more productive, then we don’t know how they will utilize the time.”

So, given the value enterprise search brings to employees and the organization, have you ever wondered why it has gotten such a bad rap?

It is because implementations didn’t meet users’ or leadership’s expectations. So, a lot of money was spent, lots of resources were used, and the results were awful. In my opinion, the primary reasons for this are:

  1. Monolithic Implementations:  Utilizing “ERP-like” integrated platforms that are heavily customized frequently take six months or more to implement, and they are painful to upgrade.  SharePoint 2010 had some of this, and so do other ERP-Like platforms.
  2. High Total Cost of Ownership:  Inherent in the “all or nothing” approach of implementing a closed platform is the purchase and implementation of features that aren’t used. In some cases, this has driven TCO up to 2-5x higher than necessary.
  3. Lack of Flexibility Large integrated platforms are expensive and difficult to deploy and upgrade, slowing the responsiveness of information technology. This creates user dissatisfaction throughout the organization, damages the reputation of the information technology team, and makes “search” a bad word.

In the past five or so years a lot has changed, and now there is a better way of addressing enterprise, or any internal, search initiative. Our recommendations to customers are:

  • Take Advantage of Your Existing Infrastructure Whether it is O365, Azure Search, Elasticsearch, or SharePoint Server, use technology that works within your current environment. You don’t need another platform.
  • Use a Modular Solution Choose, implement, and pay for only what you need.  This provides the flexibility to implement what you need now, add to it over time, or upgrade certain components without upgrading everything.
  • Protect Your Investment:  With the sunsetting of GSA, many organizations face costly rip and replace efforts. Your investment should be protected. For example, if you start with SharePoint and migrate to Elasticsearch, then you shouldn’t have to start from scratch. You should also be able to unify users’ search results.
  • Add Intelligence: Take advantage of AI platforms such as those from Microsoft, Google and the Rasa Stack.  A new wave of Artificial Intelligence technology offers the potential to dramatically transform the way you access and use your information.
  • Resist customizing: Don’t build something that you can buy out of the box (OOTB). Take a software-driven approach first, and if you cannot find what you need, then think seriously about a custom solution versus finding a vendor who would do it as an OOTB software package.

Leave a Reply

Your email address will not be published. Required fields are marked *

For security, use of Google's reCAPTCHA service is required which is subject to the Google Privacy Policy and Terms of Use.

If you agree to these terms, please click here.