The Interview and the Data

17 02 2011

Had a conversation yesterday that summed up the interviewer’s dilemma rather well. They said:

“We put an enormous amount of effort into interviewing the highly intelligent, bright young things who apply for jobs at our firm; and we are rigorous in only taking those who pass with flying colours. Yet somehow we end up with a large percentage of people who simply can’t apply themselves to the very particular demands of the work we do.”

So what is going on here? It would be easy to focus on young graduates being taught to present themselves in a desirable light and so forth, but actually that simply ducks the fundamental question: what is the job interview for in the first place?

Perhaps it is easier to say first what the interview is NOT for: how a person presents themselves at interview gives you NO positive guidance whatsoever as to how they will discharge their duties in a job (unless you are hiring them as an interviewer!). I say positive guidance; negative observations (poor interpersonal skills, poor personal presentation etc) may well transfer to real life situations. But ask yourself – does that matter for this job? The person who interviews awkwardly may be the person who can deliver the results in the role. What I am mostly getting at is that “being personable”, “exuding gravitas”, “good answer to a difficult question” – none of these actually tell you how a person will discharge actual duties in a real job. Unfortunate but true.

Submitting to interview is a bit like undergoing trial by ordeal in medieval times. Uncomfortable to watch, can be entertaining for (heartless) observers, and can even be survived (occasionally) by the subject; but ultimately unlikely to distinguish accurately between the innocent and the villain. So why on earth do we do it, and why do I still think it is a critical part of the recruitment process? Here it is:

The interview allows us (as the panel) to push down hard, but fairly, into the data, and for you (the interviewee) to demonstrate where the reality lies.

Which begs the question: what data?

First there is track record. This can be a bit thin in the case of a new graduate, but the graduate who has never done anything except attend class must accept they are at a disadvantage of their own making. It hardly needs to be said that we are all tempted to say “achieved 300% growth in sales over 15 month period” when what we actually mean is “I joined the team when they were on a bit of a roll”. The latter case may still mean “…and as a result I learnt a great deal and had the experience of being part of a top sales team”, which may be extremely valuable experience. It just isn’t precisely what is implied by the former statement.

Track record needs to be unpicked. The rule of thumb in all interviews is to ask concrete questions and then listen hard for conceptual – rather than concrete – replies. If the question is “tell us how you planned your sales programme during those 15 months”, a conceptual reply is “yes, I think planning is important / was a key part of my success / etc”. A concrete reply is “on day 3 of my time there I gathered the team for an off-site and we identified x, y and z. I decided we were going to ignore x and y for three months and focus on z instead. We never did get back to x or y.” See the difference? The concrete reply validates the claim (of course you may still have other questions: what were the long-term implications of your decision, opportunity costs, what is happening now, etc); the conceptual one undermines it.

So track record needs to be pushed hard if you want to get at reality.  “Are the claimed results real and if so, what behaviours delivered them; and therefore what behaviours might we expect to see if you work for us?”

The other kind of data is of course the result of assessment testing. It helps if you are using a test which a) allows for genuine like-for-like comparisons between individuals and which is strongly predictive of performance in b) role and in c) social and organisational context. Apart from giving someone a three month probationary trial, I only know of one tool that ticks all three of those boxes: the Birkman Method; if you know of another, please let me know.

Once you have that assessment data, you treat it just the same as the track record; you ask questions that require concrete answers and then listen hard for any conceptual ones. “This role will require you to be on the road 2 weeks out of 3; the data suggests that you might find that hard to sustain. Tell us about a role in which you have had to sustain that amount of travel.” Concrete answers can include “I worked for 9 months doing x” and “I have never done a role like that”. Conceptual answers are like “I think the key is keeping in touch with the family every day” (or whatever). Either of those concrete answers can be pursued further; the conceptual answer pretty much invites a big red cross; not only does this person not have the experience but also probably lacks self-awareness in this area.

And write, in large letters, where you can see them and the interviewee can’t, the following note-to-self: CONCRETE? Because even when you know that is what you are meant to be listening for, it is very easy to be charmed off your moorings by the accomplished interviewee. The once or twice I have been on a panel where we have allowed this to happen, we have all lived to regret it!

Advertisements

Actions

Information

One response

22 02 2011
Andy Murphy

Incisive stuff as usual Jon. This is exactly the challenge we have in our large organisation. Having sent this note, my next step is to Google “Birkman”.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s




%d bloggers like this: