If you are a business operator, you don’t actually care about ‘becoming data driven’ — you care about things like increasing profitability or decreasing churn, or some other business metric that impacts your life.

So it’s worth asking why it’s necessary to invest time and energy to ‘becoming data driven’. Xmrit is a free tool that helps you create XmR charts of your metrics; XmR charts help you identify when something in your data is worth investigating.

But why should you spend time looking at metrics in the first place?

I think it’s worthwhile to spend a little time talking about the why. Let’s face it: using data is not free — you must invest time and effort to collect and clean data before you can use it. This is drudge work. And it doesn’t get better over time: as anyone who has been involved with a sizeable sales org would know, it is a constant grind to get salespeople to enter data in a timely and reliable manner. And that’s with the sales org, mind, where there is a pre-existing culture of data use! Things tend to be worse in other business functions.

Knowing why you’re doing all of this drudge work can be useful. You’ll be more patient and more willing to invest resources in measurement if you know why you’re pursuing data in the first place.

What is Knowledge?

What is the purpose of data?

Statistician W. Edwards Deming liked to say that the purpose of data is knowledge. We shall define knowledge as ’theories or models that let you better predict the outcomes of your business actions.’

This sounds like a weird way to frame things, but the definition is quite useful: knowledge guides everything you do in your business. Whenever you make a decision to run a marketing campaign, change your sales process, or hire a software engineer, you are in effect saying: “I predict that doing this will lead to good outcomes X, Y and Z.”

Most competent business operators get better at making business decisions over time. They learn through trial and error. You should already be familiar with this process: you make some decisions, observe the outcomes of those decisions, and then reflect on them so that you may do things better in the future. But this process tends to be slow. And it is not particularly scalable: just because you’re able to reflect on your mistakes doesn’t mean everyone in your business will be able to do so.

The purpose of data is to accelerate this process. The end result is something truly wonderful: you gain a causal model of your business in your head.

What does this causal model look like? Well, you know what activities are most likely to work. You know how to increase sales, or which marketing channels work best, or which recruitment methods result in the best talent for the salary levels you’re currently able to pay. You know how to observe changes (and how to confirm this through qualitative judgment) when you have successfully introduced a change.

More importantly, you will have the ability to discover and evaluate new methods (or new channels or new opportunities or new cost cutting initiatives) for your business. You are constantly learning, and you are constantly doing better.

And you can teach these methods to everyone in your company; you no longer have to rely on the tacit, hard-won expertise that’s locked up inside your head.

Earning Knowledge Without Data

You do not need to use quantitative data to gain knowledge.

Most mid-sized companies or startups I know do NOT use data (read: they do not know how) but swear by a method of data collection called ’talking to your customers’. Talking to your customers leads to knowledge. You might already be familiar with the following cycle:

1. Some set of customers tell you that they are interested in buying your product because your product helps them solve X.
2. You come up with new feature ideas (or new marketing!) that better help them solve X.
3. You launch these features.
4. You follow up and see if new customers are helped by the new feature in future sales or customer support calls.

In truth, good business people instinctively try to ’talk to customers’. They try to do so even if they are very busy. Legendary Intel CEO Andy Grove would set aside a small amount of time each week to read customer complaints (source). You should always spend some time getting to a deep qualitative understanding of your customer if you want to be a good business operator.

But why stop at qualitative data? If you have two hands, you should use both of them. Most people default to talking to customers because that’s all they know. They do not know how to use data; nobody has ever taken them by the hand to show them an alternative way to pursue knowledge.

There is no reason to limit yourself to just one hand. You just have to be shown how.

Gaining Knowledge With Data

How do you gain knowledge with data?

The answer requires a simple reframing of your business. Think of your business as a process that consists of inputs and outputs. Every sufficiently complex process may be decomposed into multiple other processes. Each of those processes have themselves inputs and outputs.

Your job is to discover the controllable inputs for each and every process in your business.

How do you do this? Xmrit already gives you one powerful tool: the XmR chart. XmR charts tell you when a wiggling metric has successfully changed.

You will use the XmR chart in one of two ways:

1. You will make a change, and then wait to see if the XmR chart tells you if your numbers have successfully changed as a result of your actions.

2. You will observe a metric regularly, and only investigate when your XmR chart tells you that something weird is going on — perhaps due to some external factor you don’t know about.

In both cases, you will end up gaining real knowledge: that is, knowledge that some causal thing actually affects the metric you care about.

If the impact on the output metric is positive, you should find ways to do more of that thing. Run experiments; incentivise more of that behaviour.

If the impact is negative, you should modify your process to prevent that thing from occurring in the future.

And then rinse and repeat, across every output metric that you care about.

This is how you can use data to gain knowledge.

And this is how you can become great at your business.

Wrapping Up

In truth, there is nothing profound with what we’ve covered today.

Once we have defined knowledge as ’theories or models with predictive power’, we can immediately see that we what we truly care about in business is in discovering things that are causal. We want to know if a marketing activity actually increases our leads, or if a new software engineering process actually reduces bug count, or if a different recruiting process results in materially better hires.

We can do this qualitatively — by making the change, waiting for the impact, and squinting at the outcomes we observe. (And make no mistake: some people are genuinely good at this.) Or we can do this through data: by measuring some output, watching for changes, and then updating our beliefs as the result of successful (or unsuccessful!) change.

The purpose of data is knowledge. The prize at the end of all the drudge work, then, is clear: you have a path to actually understanding your business. You have a way to win.