Monitor your business’ success by determining the right performance indicators (KPIs) 🔑

A guide for small business owners and founders

Guess what? Our release date’s approaching! Yesterday, we had a team call to discuss our todo list for the Yala (our brilliant social media scheduling bot) launch. One of the new items on my list, and one of the more interesting challenges, is determining a KPI for Yala.

What is a KPI, you ask? To quote the Wikipedia article: Key performance indicators evaluate the success of an organization or of a particular activity in which it engages.

In less fancy terms, a KPI is a measure of your business’ health.

If you’re a small business owner or a startup founder, you know it’s important for you to keep a close watch on your business’ pulse. The problem is, with so many different metrics and measurements available, honing in on the right ones to track and optimize for can be hard.

I’ve already started thinking about our KPI a little while before our team call, but I couldn’t put my finger on a metric that’ll be representative enough of our business.

So I went ahead and emailed the brilliant Dean Levitt (of Teacup Analytics), who was able to teach me a methodical approach to determining KPIs, the same approach he’s used in the past at Mad Mimi. In this article, I’d like to introduce you to this approach. But first, I’d like to introduce you to Mike.

👦🏽

Mike is the owner of a neighborhood bodega. He wants to take a data-driven approach to improving his business.

To do so, Mike needs to track a certain set of metrics. Some things he might measure are:

• Inventory
• Product popularity
• Product placement
• Margins
• Wages
• Average monthly customers
• Average basket value

All of these have a significant effect on Mike’s bodega. However, which metrics are most representative of his business? How can Mike isolate those pivotal metrics and improve them?

This is where our methodical approach to determining a KPI comes in.

4 steps to determining your KPI

1. Determine your goal in relation to your customers. What would be the result of the ideal customer’s visit to Mike’s bodega? Supposing his average customer spends $10 in the store per month, Mike might want to increase that number to $11.

2. Segment and analyze your audience. List all the customers who haven’t met your goal and those who have separately. Mike will create a list of all his customers, segmented into those who spent less than $11 and those who’ve spent more at his store.

3. Find the trends. This is where it gets tricky. Mike needs to look at all of the customers who haven’t met his goal, those who’ve purchased $10 worth of items or less per month, and find common behavior patterns. For example: very few may be purchasing non-organic products; Or many might be spending less than five minutes in the store.

3½. Ensure these behaviors aren’t also common amongst high-performing customers. If a behavior isn’t unique to customers who haven’t met your goal, it’s unlikely it is indicative of that type of customer. Mike might find out some of his best customers also typically spend less than 5 minutes in his bodega, for example.

4. Extract the KPI. The performance indicator will be this secondary metric affecting these customers’ behavior. How many organic products does the bodega have in stock? How long does a customer typically stay in the store?

These are metrics that are trackable, but more importantly, leverageable. Mike is able to increase the number of organic products in his bodega or speed up the checkout process. If Mike chose the right behavior patterns, those that are unique to his lowest-spending customers, these changes should have direct effect on them, and increase his bottom line.

This process isn’t error-proof. Here are some things Mike needs to watch out for:

• Correlation vs. causation: perhaps lower-income customers don’t tend to hang around in the store. In this case, shorter visits might not be the cause of lower spend, but instead just be correlated with it.
• Characteristics vs. behaviors: the fact that some healthy-minded folks are visiting the store doesn’t mean it should stock up on organic vegetables; They might just be there for the laundry detergent.

Finding Yala’s KPI

From Mike’s bodega to our very own Yala. As I’ve mentioned before, we’re gearing up towards our launch and need to determine our KPI. Here’s what I did to do so.

Step one: determining our goal. We decided that, at this point, our goal is that every user schedules a post with Yala.

Why?

It means that people gave Yala a shot. If they went on to create additional posts, great — it means we might have a sticky product. If they didn’t, we can ask them what went wrong with that first post. Either way, achieving this goal means finding out important information about our business. Since we haven’t monetized yet, information is our most important asset!

Our goal is flexible, and it will probably change at some point (when we either achieve it or determine an even better goal). A short-term goal is okay, but it needs to remain for long enough so that we can actually measure changes.

Step two: I used Mixpanel to filter out all users who have created a post.

Step three: this is the tricky part. Thankfully, we still haven’t publicly released Yala, so our user base is still quite small. I looked at every one of them to try to determine the lowest common denominator, their essential, shared behavior pattern or patterns.

The first thing Yala does after installation is introduce herself and ask the user to connect a social account. Guess what? Most of our least-engaged users never connected that account. In addition, our engaged users can’t be engaged unless they connected an account, so this behavior is unique to our under-performing segment. Score!

So our KPI for now is percent of users who connected at least one social account. That’s a metric we can manipulate — provide more reminders, emails, easier process — to increase our goal of having most users schedule at least one post. If it turns out to have no effect on our goal, there’s proabably an even lower common denominator that I missed.

The fun part is, this KPI also makes sense. If a user hasn’t connected an account to Yala, they’re obviously not going to enjoy the timing benefit, which we know is one of the main motivators for people to use Yala. If more people experience this benefit, they’ll have a better experience with our product, and we’ll have more information about usage patterns and what we’re doing wrong.

Isn’t this a doozy?

If you enjoyed this article, please recommend so that more business owners can discover it!