So you think you’re data-driven?
If you have data, and especially if you use data in some way, you must be data-driven, right? To be sure — you need data to become data-driven — but it goes well beyond the data. It’s a mindset that will help you focus, and motivate you to achieve all that you can.
Today we have over 40+ autonomous teams in TransferWise. Some are data-driven, some have a way to go. I've recently been helping our recruiting team to take the first steps toward a bright data-driven future.
Why bother becoming data-driven?
First let’s start with why. Why should you even focus on this? Being data-driven will help you to flesh out where to put your long and short term focus. It’ll bring your mission to life and make it tangible. You, your team, and the rest of the organization will understand your mission more clearly. You’ll be driving your mission, not reacting to the rest of the organization’s latest forecast.
Teams that are data-driven will use data to understand:
- How they’re actually doing (compared to their hypotheses set before the quarter)
- Where to focus next (what should we prioritise)
- What will happen as a result (what their expected impact will be)
Teams that are doing all three are learning fast, are laser focused on the biggest opportunities, and good at forecasting their impact. Combined, these make strong teams and a powerful organization.
There are just two key ingredients:
- You need some data — most teams have that.
- You also need a clear vision/mission for your team — most teams have that too. The vision has to be problem oriented. What problems are you aiming to solve for your customers or your organization? It can’t be a checklist of features you want to build.
Your team will become an unstoppable data-driven machine when you combine the vision and the data in the right way!
It helps to look at a practical example I’ll share some ideas that the recruiting team in TransferWise is using to become more data-driven. The concepts here apply just as well to any other TransferWise team or indeed teams in other organization.
It starts with a simple table.
Start with one small table
Many teams, including our recruiting team, want to be data-driven. And they regularly use data to make decisions. For instance, Jorick who’s one of our tech recruiters, shared this example:
“On average [a portion of our recruiting cycle] is now 22 days. I saw that giving candidates no timeframe for their HackerRank test resulted in this poor performance. So I moved on with giving them a deadline of 5 working days and scheduled a call for me to talk to an applicant within 5 working days after they had applied.”
This is an amazing example — a small change in process resulted in a quadrupling of speed. Forward progress change by change like this. But we can find more of these opportunities, and implement the most important ones faster by focusing on the overall picture.
If you have a clear mission for your team, you need to answer a few questions.
- What are the handful of metrics that are the main indicators you’re successful?
- For each metric, where are you today?
- Where will you get to in the short term (+1 quarter)?
- Where will you get to in the medium term (+4 quarters)?
It’s ok that you don’t have any idea what your metrics will look like in a year. Nobody does. The point is to guess. You’ll refine your estimates when you retro every quarter so don’t worry about getting it perfect on day one.
For our recruiting team it might be something like this below (all fake metrics and #’s):
If the numbers above were real, it starts to reveal the long and short term objectives the team aspire to. The team aim to make huge strides in the next quarter on cost, and only minimally improve other metrics. In the longer-term, the team will dramatically improve on #hired/recruiter/quarter by reducing the number of hours taken to hire by 3–4X.
A simple table like this will help you more than you can imagine. You can slice these metrics by any relevant dimension (for recruiting, by office & by team) to help you prioritise. It’ll also help you explain to those you’re working with why you’re focusing on X instead of Y problem or why London vs New York.
Finally, you’ll know where you stand and be motivated by the tremendous progress you’ll make over time. You’ll probably achieve the lofty goals you set because you’re more likely to get what you measure. So that’s cool too! All that from a little table.
Be aware that — although powerful — it’s not all encompassing. Some of your team’s time will go to ‘hygiene work’ — tasks that — similar to showering — you just need to do. Mandatory reporting by regulators, helping other teams, and doing 1–1s with your team are all essential — but don’t lead directly to metric improvements.
Also be aware that your metric table must be balanced. For instance, if the recruiting team only focused on one of those metrics — say cost to hire — they could easily slash the cost by never using headhunters. But it would slow hiring tremendously with knock on effects for all other metrics. This small will automatically help balance your teams efforts.
My colleague Ilya shared some excellent thoughts on how to choose KPIs for your team.
But how do you know whether you’re data-driven?
A sign you’re becoming data-driven
Being data-driven is not just about the collecting data. It’s not just using data somehow to improve your decisions. Data driven teams talk differently because they understand the problem they’re solving more clearly. To bring it to life, here are a couple of example.
A team that has data will show a chart like this:
But so what? Why should anyone care about this chart? What story does this chart tell? A team that’s data-driven will be able to add a couple of accompanying sentences like this:
“Surprisingly, we needed to hire 44 backfill roles. This was 2X more than we expected. As a result, we’re doing X and Y next quarter. We expect to need to hire 54 next quarter and 70 in Q4”
With that additional text it’s clear how the data fits into the journey the team is on. It’s clear where they are today and where they plan to go.
Another example. The chart below helpfully describes how candidates perceived our hiring process:
Data-driven teams might add a couple of sentences like this:
“In Q2 we aimed to have 75% Agree or strongly agree, but only achieved 68%. The main challenges were in teams X and Y, and particularly in A and B offices. As a result, we’re focused on doing G and H, which we think will drive the overall score up to 82% in Q3 and 85–90% by Q4”
A team that can turn a chart into a brief story has a powerful tool that they can use to understand their own journey towards their mission, communicate it to other teams, and prioritise next steps:
The vision is clear (“we aimed…”), progress is transparent (“but only achieved…”) and well understood (“particularly in…”) and a plan is in place (“we’re focused on …in Q3”).
All that’s left is to debate about how they’ll achieve it. A simple chart becomes a powerful tool. It explains what’s important, what the team learned, and what they think they’ll need to do next. Feedback from leads or other stakeholders becomes enormously more useful when the what (metrics) and the why (vision) are clearly understood.
It’s not hard, and it pays off immediately
Teams that are data-driven use data, but don’t stop there. Being data-driven is simply adding measurement to the mission. The first step is to take a stab at making the four column, 3–5 row table above. It’ll take you an hour max.
Don’t aim for perfection at the start. Aim only to get something written down. If it’s written down it’s debatable. Debate it within the team and with your key stakeholders. You’ll naturally make the metric definitions clearer over time, the targets more realistic, and the projects better scoped. So just start.
Often, the hard part of becoming data-driven the stark clarity it provides. Some teams will avoid it because it raises some challenging questions for the team. What have we actually learned? What do we need to focus on? Where will we likely end up? Taking a data-driven approach puts these questions out there. And once asked, the team must find an answer.
Data gives you feedback as a team on what to improve, but also provides a pat on the back for things that you've nailed. Teams that become data-driven can see their progress clearly. As a result, they’re motivated to achieve even more. They learn faster, aim higher and deliver more.
Thanks to Ilya Leyrikh, Laur Läänamets, Eero Ringmäe and Dhruv Chadha for your instrumental feedback.
