// Commented out mbk Feb 2019 // //
Search this site:
Enterprise Search Blog
« NIE Newsletter

Do You Really Need a New Search Engine?

Last Updated Feb 2009

By Miles Kehoe, New Idea Engineering, Inc. - Volume 5 Number 6 - December 2008

It's hard to avoid the news that companies are trying to manage expenses as the economy slows. We've seen companies delay major software initiatives here in the US, and we have anecdotal evidence of the same kind of slowdown in Europe. But does it make sense to cut back on an expense that can make you more profitable?

Search for companies almost always falls into one of two categories: companies hope to cut expenses by enabling customer self-service and diverting customer support calls; or to directly improve the bottom line by improving marketing and selling of product via the web channel. For years, enterprise search companies and analysts have quoted figures showing the amount of time knowledge workers spend searching for information, and extrapolating to the monies that are saved by implementing great search.

There's more than just cost savings in search solutions. Martin White of Intranet Focus Ltd. suggests that companies should consider risk avoidance, not just traditional ROI, when it comes to enterprise search budgets. Consider leading search companies like Autonomy are now well-positioned for compliance, almost a pure risk management play.

We're all for implementing great new technology where there is a realistic expectation of a return. But do you really need to switch search technologies to get more bang for your buck?

Methodology not Technology

Gartner and Forrester are but two of the firms that track enterprise search companies in their annual reports. We should not be surprised that both the Gartner Magic Quadrant for Information Access and the Forrester Wave™: Enterprise Search are remarkably similar findings: consider the leaders in both reports:

Gartner Forrester
Autonomy Autonomy
Endeca Endeca
Exalead Google
Microsoft/FAST InQuira
Recommind Microsoft/FAST
Vivisimo Oracle
ZyLab Recommind
Table 1: Top-Ranked Enterprise Search Companies

Note: Gartner includes Google and Oracle in their "Challengers" category in the top "Ability to Execute" half; so when it boils down to it, Exalead, InQuira and ZyLab are about the only companies that are not leaders on both reports.

Why is this important to you?

Well, consider this: if you work for a large company, there's a high probability that you are using search technology from one of the vendors in Table 1. If your users are unhappy with your search solution, what makes you think that switching to a different vendor will make your search better?

In any given week, we may talk with a customer who is unhappy with their vendor ("Vendor X") and who wants to switch to "Vendor Y"; yet we'll also hear from a company unhappy with "Vendor Y" who is switching to "Vendor X".

Maybe the problem is not your search technology: it's the methodology.

Improve your Methodology

So you have an existing enterprise search solution - either serving intranet or customer-facing content. You had great expectations for the application, but it hasn't worked out the way you had hoped. Your users are unhappy, and complain about not being able to find the information they need. Maybe customers still call your support group for answers rather than look up answers on your web site, or your sales remained flat after the roll-out of the new search. What can you do?

First, try to understand what the problem is. Do you know your top search queries? Do your users employ the same vocabulary your content authors use? Have you tried the searches yourself, to identify the documents that come back? Have you looked for product on your site? Here's an experiment to try if your company makes a product: take a look on the label and search for the name printed on the label: can you find it from your search? One of my pet peeves is searching a vendor site for the model number on the back of one of our laptop PCs; only to find "No Hits" in their search engine. Trust me, I didn't make up the product number "fpco2150ak"!

Good reasons to keep

It works pretty well, and you can identify a way to solve the most common problems users cite. You can use the capabilities provided by your search vendor to enhance relevance by query tuning, or answer common user queries using best bets or results promotion. Your search activity review will have identified top queries, and if you use these recommendation tools or synonym/thesaurus capability, you can improve your results. Usability your biggest complain? Fund a project to update your interface to be search-best-practices compliant and easier to use and understand. Add third party solutions to extend your technology if you want to extend your capabilities without breaking the bank - you can probably add capabilities it at a far lower cost than replacing the search engine completely.

Don't toss out a technology that is already meeting 80% of your requirements.

Bad reasons to keep

Just as there are good reasons to remain with your vendor, there are reasons that should not carry much weight.

The 'status quo' is one of those bad reasons. "We've always used ‘Vendor X'".

Sometimes we hear IT departments say 'users are not complaining' or 'no one uses search'. Our guess, when we hear this, is that users have long since given up and have either switched to Google on the public web in conjunction with desktop search tools.

Even worse: "Our vendor tells us no one has better technology."

Time to Change

Sometimes change is good. New projects always have a certain air of excitement, and everyone is optimistic. And if things are not going well now, change offers the promise of better times ahead. But before you decide to change, make sure you're doing it for the right reasons. Your users are complaining, management is unhappy, and change is in the air. But is changing the technology the best solution?

Good reasons to Switch Search engines

There are a number of good reasons to cut your losses and switch to a new solution.

Maybe your current technology cannot scale. You document count has grown exponentially, and you just can't index all the content you have. A related limitation may be the raw number of queries you get: maybe it's a licensing issue, or maybe your search just can't handle the peak load. But it's just not able to handle the load. And the technology just won't support distributed servers.

Another good reason may be based on capabilities and features. Users have high expectations for search given the public web sites do it so well, and your existing technology just doesn't support the things people have come to expect. For example, you cannot use best bets/recommended links; or perhaps the document summaries are poor, or even worse, not available.

Finally, if you have licensed your existing technology on a term basis, the expiration of your license may be a good reason to consider changing vendors.

Bad Reasons to change

While change is sometimes good, there are also reason why changing your search technology may not solve the problem.

Your users are unhappy, and by extension, your management team is unhappy. But there is no budget for analyzing the problems, much less fixing them. You don't have budget for staff or outside consultants who can manage your search. You can't spend money to build expertise in-house. But you can get budget for capital expenses, and software falls into that category. Stop. Take a deep breath.

Search isn't like many enterprise software products - it's not something you install and configure and never touch again. You need periodic activity reviews and updates to your synonym list and perhaps to your relevancy algorithms. If you don't have the time and money to fix the search use now, why do you think you'll have time to take care of the new engine?

Sometimes we see "compatibility issues" between customers and vendor staff. Maybe there is a long and mixed history between your company and the current vendor, or you're just not feeling the love from your new sales person. Again, stop. Think.

I once participated in a sales call on a company when I was a young impressionable System Engineer. The first words out of the customer's mouth were ones I'll never forget: "Don't you think when two companies are talking about suing each other both sides are at fault for not trying hard enough to solve the problem?" We went on to solve the problems, and I suspect if you are creative enough you can overcome any issues you may have with your current vendor.

Finally, sometimes you'll be in the position where your vendor has obsolete the product you use and no support is available. Or maybe your vendor has gone out of business. Even these extreme cases may not be great reasons to switch vendors. Does the solution work now? Can you find workarounds to any existing problems, or third party products to enhance the capabilities? Are there consultants who know your existing product well who can help you if you run into an issue? They may not have access to code fixes, but a good consultant can help identify a workaround and keep you going.

What to do today

So the question is: What to do now?

We suggest you focus on 'smart change'. If you don't regularly review your search activity, start a two week project to do so now. Talk to your users - you don't need a huge base of people to start seeing a pattern. Bring in a consultant - ideally a vendor neutral one - who can help suggest how you can fix your existing technology and who can help you decide if it's really time to move on.