Vote Now for the KMWorld Readers' Choice Awards !

Search is an Iterative Process

This article is part of the Best Practices White Paper Enterprise Search [May 2008]

Navigating the fear, uncertainty and doubt that stalk the enterprise search waters like lone white sharks boils down to two fundamental recommendations:

1. Quit chumming the waters; and
2. Hunt one shark at a time.

Search, after all, is an iterative process, in which no wave of a wand can produce the magical solution that fits all needs at all times. Anyone selling you something different is likely throwing a bridge and some snake oil into the deal. And if you don’t believe that, try turning your sentiment analysis engine loose on the Web to reveal all those customers who have been burned by many a search vendor’s "magic pill diet." But rather than push the very riddle-based lexicon we purport to debunk, let’s get right to the meat of it.

Quit Chumming the Waters
In 2008, after dozens of years of businesses buying enterprise software and systems, it seems a tad odd to bring up the notion of proper due diligence. But that’s exactly where we find ourselves in enterprise search, where a high percentage of implementations are either delayed, fouled up or left for dead as a result of the customer and vendor not being on the same page from day one.

To be fair, customers are often in a no-win situation, from the minute they’re asked to decipher 50 different vendors with 50 different messages and terminology. Much of this is beyond their control. But when we speak of chumming the waters, we’re talking about customers who go into a review process with only half the information they need to make the best decision, hence giving these sharks (aka vendors) fertile ground from which to hunt.

Therefore, the only way to truly keep these predators at bay is to know your needs inside and out as you go into the review process; this constitutes items that are within a customer’s control. Why cloud your picture with myriad 50-cent features if 99% of them have no impact on your immediate need? It’s one thing to approach enterprise search from a strategic standpoint to ensure your vendor can grow with you over time. But it’s something entirely different to assume that a vendor’s uber engine is going to make all of your findability issues disappear over night. Instead, follow these steps:

1. Define your needs in total so that you can focus the conversations on how a vendor addresses them;
2. Prioritize the needs to identify the best starting point;
3. Ask all of the right questions, and ask them again and again until you know you have the answers; and
4. Once you have the answers, politely tell your vendor, "Terrific. Now prove it to me before I spend a single dime with you."

The delightful thing about painful due diligence is you’ll find out pretty quickly which vendor has substance and which is just looking for a quick hit. What’s more, it saves you the headache in the long run when it comes to rollout plans and expectations.

Hunt One Shark at a Time
Ask just a handful of analysts how many organizations are offering true enterprisewide search to each and every member of their enterprise, and you’re guaranteed to hear most of them say, "very few." In fact, some analysts flat out refuse to call it enterprise search, due to the connotations of the term "enterprise." Stephen Arnold of Arnold IT refers to the market as "behind-the-firewall search." In its annual magic quadrant report, Gartner labels the market as "information access."

The point being that even for something as broad sweeping as e-discovery and compliance, search is being deployed to solve specific findability needs. So why pay a vendor hundreds of thousands to millions of dollars to watch them spend two years sweeping your search problems under the rug? Two years in which NONE of those problems is being solved.

This is where we get to the heart of search being an iterative process, iterative both in the sense that you start with one project and expand, as well as begin with standard functionality and get trickier over time. In practice, this boils down to starting with your highest-priority need and hunting that shark first. Not only will it get your full focus, resulting in a faster deployment and consequent return on investment, but you’ll also glean valuable "lessons learned" that will help you in the next phase of the rollout. Conducting a pilot with five to 10 users is a good start, but it’s not enough to tell you everything you’ll need to know for a deployment that’s exponentially larger.

Perhaps it’s your engineering group, or maybe it’s corporate counsel. Both groups in most organizations have real and immediate needs that can be addressed by affordable technology that’s easy to deploy, manage and use. What’s more, they give your company an immediate return on your investment, rather than waiting several months or years to get your first dollar back from your six- or seven-figure punt.

Yes, it is critical that you understand your vendor’s scalability and extensibility. Even at just tens of thousands of dollars, you want to get the maximum return on your investment. Talk to your vendor’s current customers, ping the analysts for their views and request a pilot to confirm that everything you’ve learned holds true in your own environment. Benchmarks are fine, but the variables in search are so great that you’ll never truly know how a solution will perform until you’ve tested it yourself.


Search KMWorld

Connect