Vision planning + A/B experiments

Improving the interview scheduling experience

The problems

The problems

We know these from years of moderated research studies + customer feedback + validating with data

Discovery

Discovery

Let's do one more round of generative research

Before coming up with solutions, we looked closely at all of the data we had and put together every question we could think of to learn about their scheduling process.

We partnered with a UX Researcher who interviewed 12 employers that are in charge of scheduling interviews.

Wait, no shows aren't an issue for everyone?

The more hands-on employers were with messaging candidates, the fewer no shows they experienced

Most people don't know about our scheduling tools

Some participants were describing tools they wished existed on Indeed, not knowing that those features already exist!

Candidate availability would make life easier for complex scheduling

Later stage interviews with more people and higher level people make scheduling near impossible. If employers already had the candidate’s availability, they could confidently grab the rare open windows that work for everyone.

Vision planning

Vision planning

Let's get out of the way and encourage more human to human connection

My product partner and I believed that if we can push the scheduling process to be more conversational, both sides will be more responsive and we would have more successful interviews scheduled on Indeed.

Fix a broken experience

Fix a broken experience

Let's make rescheduling less painful

Rescheduling can be one of the most painful parts of interview scheduling. Remember all of that hassle you just went through to find a time to meet? Now you get to do it again…

Before we made this improvement, the employer would only be able to see the newly suggested times in their email. It didn't appear anywhere on Indeed for them.

I worked closely with one of our devs to bring that data into this module, as well as carry that over to the scheduling modal. This saves the employer the pain of opening two tabs and fixes a dead end flow for them.

Experiment

Experiment

Detect scheduling intent within messages

The vast majority of messages sent between employers and jobseekers are trying to find a mutual time for an interview. As we learned in research, there are quite a few employers that don’t know about Indeed’s scheduling tools which could simplify the process for them.

We worked with data scientists to build a regex that would detect words and phrases such as “tomorrow at 1” – “Tuesday afternoon” – “anytime Friday”. Once we detect scheduling intent, we created a module to nudge them to use Indeed’s scheduling tools to help employers save time and effort.

+5.31%

Employers with interview invitations

+3.21%

Employers invitations that turned into interviews

7.5%

Employers who saw this module sent invitations from it

Experiment

Experiment

Bring more awareness to our tools

This was a small experiment to help draw attention to scheduling in messaging to try to help employer avoid unnecessary back and forth.

Our long term goal was to nudge employers to share their availability once we see that they started to type something like “Are you available?”, but that was a considerable amount of work. We wanted to run this little experiment in the short term.

Test group 1
Test group 2
Test group 3

Test group 3 wins - 2.94% increase in employers with 1+ invite

The interesting thing here is that this option is the only one that mentions interview formats. That brings up a bunch of questions we need to keep digging into.

  • Do people assume they can only schedule virtual interviews?

  • If people were more aware of scheduling capabilities, would they naturally use our tools more?

  • Is saving time not a valid value prop for them?

Handoff

Handoff

Our work was just beginning and we had 2-3 years worth of experiments roughly mapped out, but the company went through some widespread re-orgs. I moved onto the messaging team and handed this work over to another designer on our scheduling platform team.