10 Reasons Why Enterprise Search Vendors Face Sales Friction
November 4, 2015
I have been watching the flow of information from companies which are in the enterprise search business. Some of these firms are ones you may recognize; for example, Elastic and LucidWorks. Others may be off your radar; for instance, Attivio, BA Insight, and Coveo. Some are essentially “hidden” outfits; for example, Diffeo.
The question I pondered when I was making my return from yet another remote location, “What is enterprise search disappeared?” (This is not a grammatical error. Disappeared is one of those jargon terms which add spice to certain government professionals’ life or the opposite.)
I jotted down a list of 10 reasons. I am not sure these are particularly humorous, but the list captures my thoughts on a 17 hour airplane flight.
- Enterprise search burned its bridges with over promising and under delivering decades ago. As a result, search is a utility. Enterprise search as an enterprise application lacks credibility.
- It is easier to talk about finding information in an enterprise and embarrassing with customers cannot locate a memo they wrote 24 hours earlier. Indexing remains the weak spot in many enterprise search systems.
- The cost of normalizing information is far greater than even the most unicorn worshiping financial wizards can stomach. Changing the oil in a La Ferrari is a deal compared to the ever escalating costs of content processing.
- Users get darned annoyed with enterprise search systems which are supposed to do everything, including squeeze the Big Data forest down to a bonsai grove on a marketer’s desk. Revolt takes the form of installing a local solution and keeping the alternative off management’s radar.
- Even the simplest solutions like an appliance requires lots of baby sitting. Free and open source solutions reduce the license fee. The other costs remain.
- Smart systems are still generally stupid and humans have to get involved to make sure the automatic processes do not return increasingly off point results.
- Specialists who can make an enterprise search system actually work are tough to find and keep on the job. Hiring a search specialists ensures one thing: Ever increasing costs for engineering support.
- The darned indexing misses the names of key customers, companies, and products. No matter how many times the system is tweaked, the notion of 85 percent accuracy guarantees 15 percent frustration. Toss in videos and images and most enterprise search systems have no way to process these file types without licensing more technology.
- Many enterprise search vendors are coding the system as they are trying to close deals. Instead of a hardened product, the vendor has delivered a demonstration which has to be configured, tuned, optimized, and updated. Translated: We are writing the software as we are billing the customer.
- The security and permissions settings require a full time person. Even then, incorrect permissions can cost a company a government contract or create legal exposure.
What’s the fix? I guarantee you that reading the recommendations of mid tier consulting firms, the analysis of frustrated academics, and failed webmasters who have morphed into search gurus will not help you out.
Check out the profiles of failed vendors at www.xenky.com/vendor-profiles. These are case examples of how search tuna salad became business health hazards.
Stephen E Arnold, November 4, 2015
Comments
2 Responses to “10 Reasons Why Enterprise Search Vendors Face Sales Friction”
-
- Subscribe to Beyond Search
Feature archive
News archive
-
All good points. What is the fix or isn’t there a fix? What does search is a utility mean?
I’m going to go out on a limb here and say it’s possible there isn’t a fix.
Good enterprise search is *hard*. Building a system that promises to cope with all the challenges you describe above, out of the box, is frankly impossible. Luckily, lots of people now realise this (sadly in many cases after buying one or more ‘magic’ search products at huge expense). The only strategy that is likely to end up in delivering something useful is to build a good search team (including both internal and external people), be realistic and open about targets and metrics, use Agile and similar methodologies to achieve incremental change, and to stop believing in magic.
If you want a good example of how this is being done, look at AstraZeneca, who have built a search team under their CTO office, are working with a good technology base (Sinequa) and are realistic about what they can achieve (which appears to be quite a lot!). We’ve also got some examples from our own experience with open source projects, where a collaborative and open process is always the best way forward.