How I automated our interview feedback process in less than 15 minutes

Brad Dunn
6 min readOct 30, 2020

--

Okay - so I’ll level with you. One area I need to get much better with is preparing for interviews. On average I can have somewhere between 15 and 27 meetings a day depending on what’s going on and because of that, shaving small amounts of time off my day has a huge impact to just how much work I can get through.

A few years back, I interviewed for a position at Google in Sydney to work on the Google Maps team as a product manager and one thing I really liked was the interview format they used, which is quite well documented. Over the years I’ve stolen bits of that process and incorporated it into the way I hire, but a big part of it is just writing up simple interview guides and then making sure the hiring committee (those who provide their assessment of the candidate) are sent to the hiring manager in a consistent format.

A consistent interview format means it’s easier to compare candidates side by side instead of relying on how good the ‘vibe’ was.

A few things stuck out to me about Google’s interview format;

  • They interviewed you in isolation (at first), and graded the candidates against 4 attributes that matter to Google (Strategic, Analytical, Googlyness, and Creative)
  • Then on each attribute, they marked you as a recommendation to (Hire, Reluctantly hire, Reluctantly don’t hire, and don’t hire)
  • These scores, with about a thousand other things, are put together in a pack, which goes to someone who looks for consistent scores and then presumably makes a decision.

What I liked about this format was that the scores were unknown to the other parties in the interview. It avoids group think (at least at the beginning).

When you interview someone with a few different peers, it’s easy to be influenced by a whole range of factors. For instance, if your boss who you want to impress is really into a candidate — it shows. So more junior people in the room, if they pick up on those signals, will tend to swing their votes more towards the hire column as they want to be seen as being on board with the boss who is clearly in love with a particular candidate.

And this got me thinking…

I was preparing for a range of interviews this week for a new graduate product analyst role at Whispir. We wanted to bring someone into the product team as we’re doing a lot of substantial number crunching and data analysis these days, and I wanted someone who could help think more scientifically about the problems.

A few minutes before one of the interviews, I went to create a quick google sheet which had some interview questions on it and then thought to myself — this could be built in Whispir Workflows!

Workflows is a new product the team have been working on at Whispir. In essence, it’s a way to automate communications in a business. If this happens, then send that message.

So I started designing up my usual interview guide in workflows to see what I could do with it. We don’t normally use the product for this kind of thing, but I thought it could be fun to try.

Creating a communications workflow is so simple, it’s not even worth going into in too much detail. You just drag things around and design a process much like you would on a whiteboard.

The one I designed looked a bit like this. You can see from the image below it just has a webform trigger, which means if someone fills in this form, fire off the communications.

So in my particular case, the web form was the interview guide with questions I wanted the hiring committee to fill in.

I started out by sending me the results of the form and secondly, sending a copy to the HR team, where I changed the content of the message.

This is what the form looked like.

In essence, I put in (much like Google did) a few attributes I wanted to score the candidates against, with radio button questions. To build the form, you just drag them onto the space on the left. (Like I said, it’s pretty easy)

Then, you publish the form, and it gives you a URL which I sent via Slack to the interview panel (In this case, people on the data science team as they were joining the interview with me)

Next, you just select who receives the message, in this case, i’ll send it to myself.

Then, you design the content of the message that gets sent — and we just use attributes from the form, to create the message. (you can see what I mean below). The purple variables are inserted and in place of them, will appear the answers the interview panel has put in.

Then I figured, why not just automate a nice SMS to the candidate to say thank-you after the interview? That could be nice.

I even added in a little wait delay on the message to give the candidate some time before they got the message.

Then, when all is said and done, once the interview panel finishes the interview, the hiring manager (me in this case) gets a nice little email with the results, the interview panel don’t find out what the others scored (so you get a much more objective assessment), the HR team gets notified, and the candidate gets an SMS.

And, if you really wanted to get fancy, if a candidate scored particularly low, say, all members of the hiring panel decided it was a unanimous ‘don’t hire’ you could add a step to the workflow to provide that feedback back to the recruiter, letting them know perhaps the candidate selection might have been a bit off.

Then we just make sure that particular message only fires under certain conditions. We do this by inserting an if condition.

I just decided, if all answers are ‘don’t hire’ then send that friendly bit of feedback to the recruiter.

This workflow was actually pretty easy to setup and build and what started as a bit of fun and a way to save me time, actually made me really reflect on how I could use workflows to build a really simple and elegant recruitment interview process.

Personally, I find it hard to deliver a good candidate experience at scale with how little time I have, so by leveraging workflows I thought it might be a good way to save me time in recruitment.

If you’d be up for trying it too, we’d love some feedback on the product too — so hit me up with any questions on twitter!

--

--

Brad Dunn
Brad Dunn

Written by Brad Dunn

Product Management Executive 🖥 Writer 📚 Tea nerd 🍵 Machine Learning Enthusiast 🤖 Physics & Psychology student @ Swinburne

No responses yet