There is a pattern in how many small to mid-sized enterprises go about researching technology applications, one that does not serve them well. As I interact with colleagues, business affiliates and professional peers, I play into this behavior unwittingly. For example, how many times have you been on the asking or answering side of this question: “My organization is planning to procure a search system this year, what systems should we be looking at?” Too often, I make a quick judgment based on what little I know about the asker and toss out a few potential candidate vendor names.
This scenario plays out frequently and now I am uncomfortable because, as a consultant and search analyst, I know that there is a lot more I need to know before offering off-handed advice to that question. Here are some ideas for questions that you should be asking first so that, when someone like me wants more context, you have ready answers.
Your first step is to survey your internal landscape and clearly document the following:
- What are the business outcomes you expect to derive from the search product, who will be using it, under what circumstances and for what purpose?
- What is the scope of the content that will be indexed for retrieval? Create a content map that explicitly illustrates: What, Where, Who, When. This means capturing what the content is in terms of document types and formats, numbers and size, and topic, and where it is being created, stored and managed. You need to know who created it, owns it, and will have access to it. Finally, it helps to document when it was created and information about retention.
- Who will be involved in product selection and evaluation, who needs to sign off at every stage of selection and procurement, who will be involved in installation and deployment, and who will maintain the system on an ongoing basis?
- What is your IT infrastructure and who controls it? If a schematic is not in place that depicts at least the portion of the computing infrastructure that will be integral to your search support, it is time to make sure one is prepared. You cannot make an informed decision about appropriate and workable search solutions without this information.
You will also be wasting the time of vendors when you seek product and licensing information if you do not have all of these issues sorted out. Much of the packaging of search products is dependent on numbers of documents or size of the corpus to be indexed, how the software will be installed, and who and how many will be accessing it. Pricing information will be vague until you have concrete content “demographics” to share with prospective vendors. You can’t even establish a budget without answering the questions above, and you need a ballpark budget figure to help narrow your choices.
So, I am resolving to be more thoughtful in my responses when queried by friends and colleagues. Before answering I will be asking you for some meaningful data in advance of reeling off a list of products. It is time for you to do some preliminary research in-house before establishing the lineup of suitors. More on the next steps, next time up.
As I was developing concepts put forth in the report Enterprise Search Markets and Applications – Capitalizing on Emerging Demand I bounced around the Internet a lot to verify information I had previously noted about products listed in the vendor directory. As I did so, evidence began to emerge about the ease with which I could resurrect an earlier retrieved bit of content. It mystified me that vendors of products to aid retrieval of content would make it so difficult to find information on their own web site. One assumption of mine has been completely debunked, that vendors would use their own search product to help site visitors discover more about their products and services. It made me wonder why they would not be showcasing the full flavor of their offerings.
The report was not written to evaluate specific products but rather to give a more holistic view of how the markets for products break down and how products themselves can be categorized. In order to do the latter, it required reading about many products with which I had no hands on experience. I wanted to understand how vendors were positioning their products, what markets they felt their products are most suited to satisfy, and what search problems were best solved with their technologies. Coming up with generalizations, trends, and differentiators was one purpose for my research. When I realized how difficult it was to dig out specifics from many vendor Web sites, I moved on, probably leaving stones unturned but time was not on my side.
Now I am going back to learn more about the problem with researching search, something complained about by a number of buyers I interviewed. Vendors are not making it easy for buyers to narrow their search for search, and shame on them. This should be a “no brainer.” If you are a vendor pushing a product that is easy to install, implement and deploy, there is no better way than to put it to work on your own site. On the other hand, if you have products that are more sophisticated in terms of offering complex retrieval by leveraging refined ontologies or rules, you had better take the time to make it work well for finding nuggets on a few hundred pages of your Web site.
I am going to be writing more about this because the deeper I dig, the more interesting the results. For starters, of the first 28 vendor on my list, twelve have no site search. Of those that do, several use a third-party search engine, not their own. One major vendor’s search result count displayed nearly a hundred records that matched the search while also displaying the breakdown of records by category. The trouble was the category numbers totaled less than 20. Hmmm!
Perhaps the trouble in “searchland” is that no one wants to take the time to implement, deploy and maintain search to satisfy the user. I keep saying, “it’s not the technology; it’s the thought and skill that goes into the back room implementation.” Or is it? Stay tuned.
Amidst post Gilbane San Francisco business I have been reading what everyone else has been writing about search the past couple of months. While there continues to be much speculation and gossip about the Microsoft acquisition of FAST, and which companies may soon be absorbed into larger entities, there also continues to be interesting activity among the mid-tier and start-up search vendors. Meanwhile, I advise those who aspire to acquire a search solution for “behind the firewall,” don’t wait for the “big players” to come up with the definitive solution to all your search needs because it will never happen. I’m in good company with other analysts who advise moving on with point search solutions for specific business needs. You will save money, and time because many of the new products are optimized for rapid deployment, in weeks or months, not years.
If you check out my new research report, Enterprise Search Markets and Applications; Capitalizing on Emerging Demand, June, 2008, you will find a directory to companies offering search solutions with choices for what Steve Arnold refers to as “beyond search.” Deep test drives of many of these products can be found in his report, as well. Meanwhile, new releases of products listed, and new products both continue to be announced. ISYS, Coveo and Expert System (Cogito) have brought new offerings to market in the last month and Collexis, a relative newcomer, is drawing attention to itself by demonstrating its products at numerous meetings this year.
So, keeping reading and checking out the possibilities. While you are at it, be sure to put the Gilbane Boston Conference on your calendar for December 3 – 4. We are all busy rounding out the program right now.
I am particularly interested in hearing from those of you who have participated in the selection of a search product in the past two years, implementing or deploying a system anywhere within your own enterprise. Please consider sending me a brief proposal for a presentation at the conference. For your effort, you will get to attend all the conference sessions, as well as help the audience with the needed reality checks on what it takes to conduct a selection process and follow through with implementation. I particularly want you to share your learning experiences: the good, the frustrating, and the lessons you have accrued. Professional speaking experience is not required – we want stories. [You’ll find my email on the “Contacts” page of the Gilbane site and you should also look at the speakers guidelines for additional information.]