The right mindset is critical for growing your business. If you learn something from every change — whether it’s positive or negative — you will know exactly how to identify and capture new opportunities for growth.

Most people hate change. Some actively avoid changes to their lives because they fear a bad outcome. For them, stable mediocrity is better than the risk of failing.

As an entrepreneur, you’re more likely to be open to change. That’s what makes entrepreneurs special. But do you truly embrace change and make a point of learning from every change? That’s the key to the Growth Mindset. Growth comes from continually evolving your business to capture more and more opportunities.

As you learn more about the Growth Mindset, remember that the goal is always to be directionally correct, not precise. Financial reports require precision, but analytical thinking is about making sure you’re headed in the right direction. In this context, precision is time-consuming, impractical, and — fortunately — unnecessary for making well-informed decisions.

The Growth Mindset

If you have even the slightest aversion to change, start with this reframe: Every change is an experiment. Whether it succeeds or fails, you will learn something.

Thinking in terms of experiments removes emotion because it switches your mindset to that of an “interested observer.” The fear of failure is offset by the positive outcome of learning something new about your business.

The Growth Mindset has four steps: understand your business’s baseline performance, run an experiment, measure the results, and make adjustments to grow your business. This will result in a new baseline, and the cycle repeats from there. Let’s drill into each of these steps.

Understanding Baseline Performance

This is why it’s important to know your numbers. You need to know some key metrics so you understand how your business is performing overall, but also at a more granular level by category, product, time of day, day of week, and so on. For example, if you want to improve customer loyalty, you need to know the baseline metrics for how loyal your customers currently are.

Said another way, you can only manage what you measure. So, if you want to grow sales in the afternoons, you have to start by knowing your current sales numbers in the afternoons. If you don’t have the right tools to track this, find a better business intelligence tool.

Before moving on to the next steps, it’s important to differentiate this approach from relying on “gut feel.” You obviously know your business well and may be able to determine that sales improve in the afternoons just by observation. However, this approach ignores the richer, deeper, and unbiased insights that analytics tools provide. Both are needed.

Running Experiments

Again, every change is an experiment. Not every experiment is worth monitoring and measuring, but there are endless experiments you could run. Let’s start with the two types: planned and unplanned.

Planned experiments are the ones you choose to run. They are typically based on a decision you made to improve your business. Some examples are introducing or discontinuing a product, raising a product’s price, offering a promotion, changing store hours, launching a loyalty program, adjusting staff levels, and so on.

By planning an experiment, you can proactively capture the baseline metrics and define the success metrics (e.g., a 10% increase in sales of Product A within 90 days). This is the ideal type of experiment because they’re controlled.

The other type is unplanned experiments, which are also called “natural experiments.” These are changes that are “thrust upon you” rather than being thoughtfully planned out. They’re often a great source of insight, but also a great source of stress without the “every change is an experiment” mindset.

COVID-19 was a long series of unplanned experiments for your business. You would never choose to run these types of experiments, but it might have been the best opportunity you’ve ever had to learn more about your business.

For example, you might have tried delivery services for the first time, or relied more on online/mobile ordering. Or added a surcharge to each transaction. Or tried a different supplier or alternative products due to supply chain issues. There was so much disruption that it forced changes in almost every aspect of your business.

Some non-COVID examples of unplanned experiments include understaffed or overstaffed shifts, a new competitor, an existing competitor that closes, a shift in traffic patterns (street traffic or foot traffic in walkable areas), etc.

These types of experiments are less ideal because you’re not in control, but can be much more insightful because they are experiments you’d probably never run on purpose. You can compensate for the lack of control by retroactively determining the baseline performance and success metrics.

Measuring the Results

Most experiments will have obvious success metrics, like an increase in sales, order size, repeat customers, etc. You should typically choose more than one metric so you can assess the results from different perspectives.

For example, let’s say you launch a “BOGO Mondays” promotion to increase sales on your slowest day of the week. Sales growth would be a good metric, but so would profit margin. If you did not include profit margin in this analysis, you would likely find this to be a very positive change to your business — but potentially a very bad decision because of its negative impact on profit margin. Similarly, you should compare the change in these metrics on Mondays against the other days of the week to see the relative change. If the sales drop the other days of the week because everyone visits on Mondays to take advantage of the promotion, that’s another negative consequence you need to consider.

When you choose your metrics, make sure the measurements are statistically valid. For example, if you’re raising prices, you cannot compare the sales from the day before and the day after the price increase to see if it has the desired effect. There are too many variables and too little data to draw any conclusions in that short time frame.

Here are some factors to keep in mind when measuring results:

  • Quantity of Data Points. The natural variability in product sales, customer visits, etc., tends to average out over longer periods of time. Most changes should be measured over a few months rather than days or weeks.
  • Seasonality. Most businesses have a seasonal trend that needs to be removed from the success metrics. If you don’t, any change could look great heading into a busy season like Christmas. You can run planned experiments during a time of year that has little seasonality for your business, or try to estimate the seasonal component of each change.
  • Other Biases. Seasonality is one type of bias in the data, but it’s important to identify others. For example, if a competitor opened (or closed) during an experiment, is it possible that external change biased the results of your experiment?

But what about changes that are not measurable or that primarily impact intangibles? For example, what if the goal is to improve the business’s reputation, increase employee morale, or improve customer satisfaction?

The really long answer is to read the book “How to Measure Anything: Finding the Value of Intangibles in Business.” The short answer is that anything can be measured. It just might take more thought to determine the best (but perhaps imperfect) way to measure the outcome.

For example, customer satisfaction can seem subjective and difficult to measure accurately. So think about what satisfied customers do that dissatisfied customers do not. With this thought process, Repeat Visits seems like a reasonably good metric. The change in your Google rating would be another. In this situation, sales growth isn’t a great metric because too many factors influence it, but it’s a good secondary metric. You could also collect a new metric like Net Promoter Score for this purpose.

Choosing multiple metrics like this helps improve the measurement’s overall accuracy. This concept is similar to how GPS identifies your phone’s location. A single GPS satellite cannot accurately identify your location, but the combination of four satellites can be extremely accurate. So, look at an experiment from multiple perspectives to see if they all converge on a clear answer.

Just keep in mind that you should discard the results of an experiment if you don’t trust the measurement of it. All data is imperfect, but you don’t want to make a flawed decision because of a deeply flawed measurement of the experiment’s results. It’s better to discard the experiment completely and move on to the next one.

Making Adjustments

All this effort is wasted if you don’t act on it or at least learn from it. Even if the results are inconclusive, you at least learned that the experiment did not identify an important lever for growth. Move on to other ideas.

Some results aren’t immediately actionable. For example, if you raise the price of a product and customers show low price sensitivity, that doesn’t mean you should immediately raise the price again. Instead, you now know your next price increase can be a little sooner and/or a little bigger than normal.

Establishing a New Baseline

Every adjustment will impact your baseline performance, so you want to avoid overlapping experiments if possible. Testing too many changes at once may invalidate all of this analysis because it’s difficult to separate the impact of each change. But as soon as you conclude one experiment, have the next one ready to go so you are continually experimenting to find new opportunities for growth.

Next Steps

Shifting into the Growth Mindset may take a while, so be patient. Start by measuring the impact of one or two big changes to become more comfortable with the process — including which metrics to use, how long to run the experiment, etc. After a while, this will become second nature.

Also, take a look at the next several topics to get more specific guidance on raising prices, evaluating your store hours, and so on. They each provide more details on running experiments to learn more about these specific aspects of your business.

Next Up: Raising Prices: 5 Lessons Learned Using a Data-Driven Approach