A value-added service in many public search engines has been auto-complete and pre-search suggested searches for searchers. You go to Google or Bing and begin typing a query, for example, “bus”, and Google’s auto-complete suggests “bus times,” “business links,” “business for sale,” and “bus timetable.” This feature is very helpful in several aspects. By displaying alternative word forms, the researcher has the opportunity to discover a better formulated version of his or her query. The suggestions might be words that are directly related to the searcher’s intent but which also serve to expand or refine the user’s search. Auto-complete in mobile devices, where the logic seems to be driven by an app in the smartphone itself, such as in my iPhone, can be quite annoying because frequently the suggested terms are wholly unrelated to my search intent and I experience it as low value-add, in fact, a nuisance. But I’ve adapted my search workflow to the iPhone.

Some best practices from the professional world
Recently several tax, legal, and regulatory web-based research products across Wolters Kluwer have added auto-complete and pre-search suggested searches to the delight of customers. For example, today a customer of Wolters Kluwer Spain’s La Ley Digital can log into the system, and begin typing the letter “c.” The system immediately proposes two documents, the Spanish Constitution and a Spanish law, as well as a long list of search terms that start with the letter “c”. La Ley’s auto-complete and pre-search suggested searches thus can help a customer go directly from entering one or two letters and get specifically to a document. Now La Ley Digital is a research product for lawyers. Unlike Google, which serves literally the world, La Ley Digital’s auto-complete and pre-search suggested searches suggest entries that are relevant only to legal practitioners. La Ley goes a step further than Google by automatically adding semantically-related synonyms so that any expansion of the pre-search suggested searches are expanded in a way that matches the user’s intent. That’s how I wish my iPhone search would work, by the way. Another impressive feature of La Ley’s implementation is that even after you select the first search term, La Ley performs auto-complete and generates additional pre-search suggested searches based on the term selected in the prior operation. It goes on and on until there are no more terms for the system to suggest or the user decides that the terms best representing his or her intent have been found. I simply must say “congratulations” to the team at Wolters Kluwer Spain in achieving what I must call the “right way” to implement auto-complete and pre-search suggested searches.

Wolters Kluwer Tax & Accounting’s IntelliConnect also recently released auto-complete and pre-search suggested searches. An interesting aspect about IntelliConnect is that it offers specialized content for specific tax, legal, and regulatory domains, such as tax, securities, labor, etc., depending on what the customer has purchased. If I type “con” IntelliConnect will suggest both “concentration risk capital component” that sounds like it comes from the securities world, and “Connecticut Cigarette Tax,” clearly a word from the tax and accounting perspective. So no matter what I subscribe to in IntelliConnect, I can experience auto-complete and pre-search suggested searches in a way that is relevant to my areas of practice.

What do you think?
Auto-complete and pres-search suggested searches raise some interesting questions and I am curious about the opinion of others:

1. Should a suggested search term based on auto-complete or pres-search suggested searches ever lead to a zero-results search? Does the existence of a recommended term also suggest that there is a document to match it? Should the universe of suggested search be limited according to the content that has actually been selected?
2. Should customers have the ability to add their own auto-complete and pre-search suggested search entries when they do not appear in the list? If yes, should it be possible for all other users of the system to have access to that customer’s suggestion?
3. Is there is a way for us to personalize the universe of auto-complete entries and pre-search suggested searches according to the practice areas that are relevant to customers? For example, if I am an IntelliConnect customer and I am interested only in securities content, even though I can access any practice area on IntelliConnect, should there be a way for me to exclude from auto-complete and pre-search suggested searches anything not related to securities?

I am proud that IntelliConnect, La Ley Digital, and Wolters Kluwer Belgium’s Jura all offer auto-complete and pre-search suggested searches to help end users get more productivity out of these research products. You will see this feature appearing in many more products. Do you have any thoughts about the questions noted above? Do you have any thoughts to share based on your use of auto-complete and pre-search suggested searches? I would like to hear them.