Eliminate Lost Sales and Bad Customer Experiences

The Limitations of Keyword Search

Web self-service has become the preferred method for consumers to buy products and get services, and keyword search has become the “de facto” self-service information-access method, thanks to the ubiquitous search box. However, as B2C interactions converge across sales, marketing, and service, and across simple informational queries to complex queries on problem resolution and sales advice, the limitations of keyword search starts to show.

Many businesses provide the ubiquitous search box as the only mode of web self-service, taking a “one size fits all” approach. While the search box has caught on with consumers due to its simplicity, simple keyword search often results in search overload with lots of irrelevant hits or dead-ends, where there are no hits whatsoever.

Keyword search puts the search burden on consumers, requiring them to try and try until they come up with the right phrase that will unlock the answer.

Even leading search engine providers ask consumers to anticipate keywords that would be in the answer to increase the likelihood of getting the answer. Search tips on a popular site recommend that users should “Think how the page you are looking for will be written… Use the words that are most likely to appear on the page.” In other words, consumers need to know what the answer is or divine how an answer might have been authored in order to find the answer! The same applies to customer-facing personnel such as contact center agents and sales representatives when they search for answers under pressure in sales and service situations.

While writing this article, we searched for “unproductive information search syndrome” in all the leading search engines and found hardly anything related to the intent of the search. Articles that showed up on all leading search engines on the first page covered completely unrelated topics like unproductive cough, retching, procrastination, OCD, and fatigue. Searching for the “phenomenon of unproductive information search” got first-page hits on topics like desertification, the US space program, mind control, and petropolitics. In fact, the search engines were even unable to distinguish between “search” and “research,” and sometimes even got fooled by ads on web pages with those keywords!

How to Make Search More Effective: 6 Steps

How can businesses make the simple search box or keyword search more effective in enabling consumers and customer-facing staff alike to easily find information for better customer experience and increased sales? The following six steps will help.

1. Make keyword search more intelligent

Auto-filling, spelling clarification, synonym recognition, and natural language processing can help make basic keyword search more intelligent. Natural language processing can tell the difference between “what” and “how,” and figure out the intent, while they might simply be “noise” to standard search engines. It also understands synonyms and can differentiate between homonyms (e.g., a financial services “bank” and a river “bank”).

2. Go beyond keyword search

B2C web self-service searches are starting to converge across topics like product information, product recommendation, comparison, “how to” queries, problem resolution, etc. These topics span marketing, sales, and service functions, and range from the simple to the complex. While keyword search is suitable for some basic searches, other queries are better handled by topic-tree browsing, FAQs, guided help, or hybrid navigation across these access paths. Force-fitting keyword search on all these queries will result in poor search experiences, customer defections, and lost sales.

3. Bind them all behind a search-fronted “GPS”

While the alternative access methods mentioned above are powerful, they are also proprietary and hard to use. Here’s where a search-fronted GPS approach can help! Over the last few years, consumers have become very familiar with GPS devices. When the automobile driver enters the destination, the GPS device presents multiple ways to get to the destination—freeways, highways, side streets, and everything in between. Drivers are then able to pick the route of their choice, based on their preferences and driving styles. For instance, a driver may not like to drive on freeways and choose surface streets or choose to go on surface streets initially and take the freeway at a later point for various reasons. Moreover, the GPS guides the driver step by step to the destination, recalculating the route along the way if the driver takes a route that was not originally selected. The exact same paradigm can be applied to the search.

eGain Multisearch™, a unique capability in eGain’s knowledge base solution for agents and web self-service, presents a simple search box, where users type in what they are looking for—a product they want to buy, a problem they need to solve, a procedure they want to know about, etc. This is analogous to entering the destination into the GPS. The system then presents multiple access routes to the answer—keyword and natural language search, folder tree navigation, step-by-step guided help sessions (similar to the GPS’ step-by-step driving instructions), actual knowledge base articles, and even chatbot-assisted help. The user can then pick one of those routes to get to the answer, and even go from a folder tree to a guided help option later in the search process. This “multisearch” GPS-like paradigm enables users to find information easily, flexibly, and intuitively, instead of trying to figure out the route from a printed map, which might feel like a search overload of raw (often unrelated) information, or a dead-end if the driver can’t even find the destination on the map!

4. Widen the knowledge net

While the contact center and web self-service knowledge base might cover frequently asked questions and perhaps more in savvy organizations, additional information may reside in knowledge bases sitting elsewhere in the enterprise. Moreover, “power users” of products contribute a great deal of useful information to online forums and social networks, whether these forums are your own or run by third parties. You will increase the likelihood of helping the consumer find answers by federating search results from these multiple knowledge sources and presenting them in one place while identifying the sources in the “hits.” This allows consumers to infer trust levels of the presented answers and use the information appropriately.

5. Prescribe search paths

Providing profiled knowledge access paths is particularly useful when you have lots of novice customer service agents, when they have many topics and products to cover, or when you are in a highly-regulated industry. The same concept can be extended to consumer self-service search as well through factors such as their self-selected search preference settings.

A leading computer manufacturer and the small business banking division of an international bank insist that novice agents use guided help in handling customer queries. This is analogous to a driver getting to a new town and operating a rental car—he’s better off with a GPS device rather than a printed map. Likewise, a bus driver may have to follow a specific route, dictated by the public transportation system, just like government regulation or organizational best practice might dictate what questions should and shouldn’t be asked of consumers, and how the agent-assisted or self-service search interaction should progress.

6. Learn and optimize

Like anything else, search paths and knowledge base content can be optimized for effectiveness through continuous learning. For instance, answers (knowledge base articles in customer contact center parlance) that effectively dispose customer queries (this can be measured or inferred) can be matched to the query text and elevated to the top of search results, when the same query originates from other users. Question matching engines can help here as well, moving relevant answers to the top when there’s a one-to-one match between a knowledge base article and a consumer query. Adaptive knowledge management systems identify ineffective answers and trigger content creation and refinement tasks to make sure that the knowledge base covers all common consumer searches and it is effective in handling the queries.

Final word

Search friction can result in lost sales and poor customer experiences. Expanding the power of the simple search box with natural language processing, intent inference, multiple access methods, GPS-style navigation, and knowledge federation can help unleash search-powered value in the form of superior experiences for the consumer and topline revenue and bottom-line savings for the business!