Which KPIs Matter Most?
It’s this last metric – Churn – that is almost universally considered the most important metric to improve (see David Skok’s excellent analysis here). To demonstrate that churn is, in some situations, perhaps not the right place to start when looking to improve business results, I’ll review the following:
- The LTV associated with improving each metric.
- The impact of time on when benefits are realized.
- The relative improvements you can expect on each metric.
- Other factors to consider.
To make the comparison easier, I’m going to start with David’s churn number, while estimating ARPU and conversion. For us, that means churn at 5%, ARPU at $50, and conversion at 15%.
Let’s start off by assuming we can improve each KPI by the 50% that David uses for churn improvements (from 5% to 2.5%). This means ARPU increases from $50 to $75, while conversion goes from 15% to 22.5%. If we look at the LTV impact of each change, we see the following:
So far so good. On an individual basis, reducing churn by 50% has the same impact as improving both ARPU and conversion by 50%. But a lifetime is a long time to wait, though. In this example, let’s plot the impact of each metric on a monthly basis.
While the overall impact of churn is the largest, you have to be patient. Based on these assumptions, you have to wait between 3 and 4 years before the impact from churn exceeds the impact from increasing ARPU and Conversion. For instance, after two years, you’d have received a 250% higher return from ARPU/Conversion than you would from Churn.
Up until now, we’ve assumed that we can impact each KPI by a given amount. We should relax this assumption for two reasons:
- Every company will be performing differently on a relative basis, influencing which KPI represents their lowest-hanging fruit.
- KPIs earlier in the customer life-cycle tend to be easier to impact. In general, this means that conversion > ARPU > churn.
So let’s see what happens when we adjust each KPI, taking into account the relative ease with which we can do this (hypothetically). Let’s improve conversion by 40%, ARPU by 30%, and churn by 20%.
Every company must establish baselines for their KPIs in order to understand the likely impact of optimization programs. It may be realistic to improve churn even more than ARPU and conversion; it just depends on your (relative) starting point. For relatively well-managed companies, reducing churn (or improving retention) is less simple (and more expensive) than generally understood. I’ve seen great companies hammer away at churn/retention numbers for months, even years, without making a dent, once the easy improvements were made.
There are additional factors that should be understood before embarking on a churn/retention improvement pogrom. Here are just a few:
- Significant costs. These might be direct costs – salespeople, customer support, your product team. They can also be opportunity costs – for example, every dollar spent on reducing churn is a dollar that could have been spent on the immediate benefits of customer acquisition.
- Number of customers. I’ve seen fixed cost components of churn/retention programs average 30%. As the size of the customer base increases, these costs are spread over a greater number of customers, improving unit economics.
- Acquisition economics. Your willingness to invest in churn/retention improvements depends in part on your costs to acquire new customers. Generally speaking, as your acquisition costs increase (and/or as market growth slows), retaining existing customers becomes more important.
- Stay/go impact. How much impact do you have on the customer’s stay/go decision? If you have a relatively low impact on that decision (e.g., if your customers are small and often go out of business), churn/retention strategies may not work for you – focus instead on improving your other KPIs.
- Lag factors. How long it takes to see significant impact on your metrics is an important consideration. Conversion testing can yield results in days/weeks; ARPU generally takes months. Churn/retention strategies can take years to get right / reach full impact. Make sure you build these ramp up times into your decision process.
This analysis attempts to show that the usual recommendation of focusing on churn as the most important KPI is overly broad. A combination of general and company-specific factors need to be accounted for prior to deciding upon which metrics to focus.
This analysis assumes that a company focuses on one KPI at a time. While it’s possible to focus on multiple KPIs concurrently, in my experience this typically leads to sub-optimization for all KPIs, especially for smaller companies with less experience and more limited room for error.
For a company unsure where to start, I’d suggest the following steps:
- Accurately measure your current and historical KPIs.
- Compare your KPIs to industry averages, establishing a baseline.
- Estimate direct & opportunity costs associated with KPI optimization programs.
- Model out the expected impact from each program and prioritize accordingly.