Setting input goals and metrics
What is the goal of this thought piece and why

This article aims to teach you how to identify controllable input metrics that can guide your teams' focus and strategy.

An input is something we control (how many people reach out to in a sales team) versus an output that we don't control ($ revenue from that outreach).

More often than not, we measure success by the output we don't control, leading to bad strategies that focus on the wrong behaviours and misdiagnose failure confusing bad outputs as bad inputs.

A useful analogy

A great analogy to explain this concept is from a book called 'Loonshots' by Safi Bahcall - the concept of a hypothetical lottery.

Imagine you are given the opportunity to enter a lottery where you have a 90% chance of winning 10 million dollars and it costs you $10. Would you enter that lottery?

Of course! It is a no-brainer. But now say you enter that lottery, and you lose. Would you enter it again? And what if you enter again and then lose one more time?

If you focus on the outputs out of your control (for example revenue), then you aren't really evaluating if how you are working is right. In our lottery example, you are unlucky, but the input you control (the choice to enter the lottery) is absolutely the right choice. In business, we want to figure out what inputs we can control that over time will lead to success, not obsess and measure success by the outputs we don't control such as revenue.

Overview of the steps for identifying input metrics

The steps are as follows:

Step one: define the outputs that matter

We begin by defining the outputs we want, such as revenue. These are likely already how your targets are set and what the team's key performance indicators ('KPIs') are.

Make a list of all the outputs your team is accountable for delivering and the key performance indicators for these outputs.

For example, if I were running a product team, I would have some of the following outputs as part of my responsibilities:

Step two: identify your output drivers

Next, we use the question: 'what are the factors that drive this output?' to identify what relevant drivers we can focus on improving to improve the output over the long term.

We'll use a training product and customer experience as an example of how we do this. For any professional training, 'what are the factors that drive the customer experience?': 

We find it hard to believe that in 10 years' time, any customer is going to turn around (with respect to any of the factors above) and say: "I love how relevant your training is to my life, but could you make it less relevant?". In other words, if we improve how relevant our training is to our customer, we will improve the overall customer experience score.

For each of your outputs identified in step one, you want to identify the drivers.

Step three: identify your inputs for each driver

Once you have identified your drivers, we can begin brainstorming what inputs can be controlled that influence that driver. Ask: 'what are the things within my control that I can do to increase this driver?'.

If we take one of the drivers from above, how convenient professional training is, here are some of the inputs we can control that we assume improves convenience

These are just a few examples, there are many inputs you might control for just improving convenience. Don't expect these to be static, your inputs should evolve as you learn what actually increases convenience/your relevant driver of interest.

Optional step: prioritise inputs

If you have a large number of inputs identified, you might want to put up a 2x2 grid. On one axis, what is the resource cost to improve this input? On the other axis, what is the impact of this input on achieving our desired outcome? 

You want to focus on inputs that cost little to improve but that have a large impact on your desired outcome.

For example, we might determine that creating templates takes us little time, and costs no money (low resource cost) but has a large impact on the convenience of our training. A mobile app however is high resource cost and possibly not desired at all.

Step four: turn your inputs into metrics

Once you have identified your controllable inputs, turn them into metrics.

For example, if we were turning the introduction of templates into a relevant metric, we would use: % of taught content accompanied by a template for application at work.

Our assumption, the greater % of the taught content that has a template for application, the more convenient it will be apply the learnings and therefore this will not only drive convenience up but relevance (application of content at work) and therefore overall customer experience.

We would seek to measure trends between % templates, convenience and customer experience to be sure that this is true.

Step five: develop your data pyramid

Our final step once we have turned our inputs into metrics is to document them as a data pyramid. In the top box, we have the output, below the drivers, and below that the input metrics for each driver.