Can’t We Just Talk?

Online searching is a conscious and targeted activity. You engage online search providers such as Google, eBay,, and GO1 because you have specific information needs. The information that you gather from the search could be prerequisite to getting a task done or simply part of your everyday learning. The thing that makes search interesting is that needs come in different levels of complexity.


There are information needs which can be satisfied through only a few searches within a short period of time. They are primarily routine needs such as finding out the meaning of a word or the phone number of a restaurant on Google, the recipe for a dish from, or simply to bag a bargain on eBay.

On the other hand, we have complex information needs which play a big part in many of today’s important decision making. Examples include looking for your first home on, finding employment on SEEK, and updating your skills and knowledge on GO1 to keep abreast of changes in your profession. The answers to complex needs are often beyond the reach of a routine search or two. It takes months, for instance, for a person to research for opportunities online and offline before landing an actual job. This requires a person to conduct many searches with online providers, attend networking events and turn up for many interviews over the lifetime of job seeking.

“…this is what I typed, here’s what I meant…” [1]

The more complex the needs are, the harder it is for users to articulate their intent completely as queries. This gap between the actual intent versus what the users can or are willing to express as queries contributes to results which are not optimised for the users’ needs. We will discuss some examples in online job seeking and discuss ways we can plug this gap.

Why queries alone may not be enough?

Assume that the need of a user is to secure employment as a project manager in the ICT sector in Melbourne immediately. Today, users’ best attempt at expressing intent involves the obvious such as role titles, industries and locations. With such queries, even if the search is able to return all project manager opportunities in the requested industry and location and nothing else, the job is really only half done.

“Search in the last decade has moved from ‘give me what I said’ to ‘give me what I want’.” [2]

Why?”, you may ask. In this scenario, it is true that the results are spot on and the search appears successful. This however is a very traditional academic view of search. The reason is these results are relevant only in relation to a small part of the need which happens to be more easily articulated using words. This perspective is void of any concerns about the success of the actual tasks, for example finding a job, based on the information which is the job ads that the search provides.

Question mark

We know that everyday, hundreds and thousands of job seekers issue exactly the same queries. But does that mean the same sets of jobs, which satisfy the explicit intent, would offer equal chances of success if applied for by the different job seekers? If this is the case, then why would hiring managers go through the expensive process of shortlisting, interviews, reference checks, etc.

The implicit things that matter beyond just the queries

In the search for employment, there are many factors that decide placements, beyond just the obvious that a person wants a job in project management in ICT. Some of these factors include how long has a person been doing project management, the kinds of projects that the person has been involved in, what are the person’s qualifications or skill gaps, questions around culture fit and so on.

“…does relevance amount to just that; the words we can express in a search box?” [3]

As discussed in a previous article on relevance, these factors are all valid part of the intent. It so happens that they are much harder to specify explicitly, if not impossible. By incorporating these factors into search, we can potentially present the jobs that the candidates are more likely to succeed in. This is in contrast to just giving the users everything that superficially looks relevant and leave them to sift through a potentially large pool of opportunities before proceeding to the next phase of their job hunting.

 Looking Out

The contexts which are important to a search for jobs that the users cannot often provide explicitly can come from many sources. The resume or CV of a person tells us the history of one’s career, the past companies, and the person’s qualifications and skills. We can also infer and make assumptions about the desire for career progression, lateral movement and company preferences from the behaviour on site of a person or the crowd of similar backgrounds at an aggregate level.

You are the query

There is of course a limit to how much we can extract and infer from data for use in search. Moreover, many of the inferences drawn about job seekers are really just educated guesses about them. Naturally, we ask the question as to why do we not just ‘talk’ to the job seekers to fill in on the things we do not know about them, the things about them that change over time and the things that we assumed wrong about them.

Imagine a world where each of us has a career companion. Someone with whom we talk about bad bosses, poor remuneration and the lack of development opportunities. Someone who knows every intimate detail about our career history, aspirations and routines at work and skill gaps.

Imagine someone who knows about the state of the economy and the health of your company and other businesses in the industry. If the companion found out that a company which you have demonstrated interest in (based on your behaviour) has recently laid off employees in roles not too dissimilar to your current one, you would be advised to reconsider your position.

Imagine this someone who would talk to you about your skill gaps or compliance needs. This advisor would make sure your next move is aligned with your career aspirations and recommend courses and other ways to plug the gaps. Once in a while, the companion would also offer you a summary of the state of the industry and the opportunities which are out there.

You will always have the option to take control of the wheels and proactively search for learning or employment opportunities. If the companion detects behaviours that deviate from your career path, they would discuss with you and update their understanding about you.

Imagine that 6 weeks before you are about to be made redundant or 3 months before you decide you have had enough, this advisor would know. This someone would do their best to find and recommend to you opportunities which you are more likely to succeed in. If you noticed that the suggestions are not aligned with what you had in mind, you discuss with them and they will learn.


Over time, the conversations will help build an understanding of the circumstances around a person’s career and the evolving learning needs. This is in addition to the queries that people still use for proactive searching, the facts that we can extract about their careers provided in CVs and learning records, and the less obvious signs which we infer from their behaviours.

Learning and career are after all a lifelong journey which is best traveled with a companion. Someone who will always act in your best interest. Imagine that this companion is an app on your smartphone. Impossible you say?


Comments are closed.

%d bloggers like this: