We’ve all seen the stats…(if you haven’t, it’s time to read up)

“It costs 6–7 times more to acquire a new customer than retain an existing one” – Bain & Company

“A 5% reduction in the customer defection rate can increase profits by 5 – 95%.” – Bain & Company

Ask yourself a simple question:
How “sticky” is your content?
Online business are concerned with how ‘sticky’ their content is (be it an article, video, review, product, etc). Or, more importantly: Are my users coming back?


 

It’s time to get REALLY serious about retention

This post outlines 3 high-level ways to approach calculating retention. In some respects each method represents an improvement over a shortcoming of the previous method. In fact, each method is valid within its domain and all can be useful.

Your goal should be to convert each new visitor to your site to a paying user (eCommerce) or a loyal habitual reader (media publisher) as quickly as possible. If your product churns too many new visitors after a single visit you have a problem because acquiring new users is much more expensive than retaining existing ones. You have a “leaky bucket”. Acquiring new users to ‘patch a leaky bucket’ is very costly, which is why the economics favor investing in increasing user retention. Because we can only improve what we measure, tracking user behavior and deciding on a retention methodology is the first step.

Every business will look at product retention through a slightly different lens – a mobile game developer lives and dies by day 1 retention while a B2B SaaS provider needs to track retention at month 18.

Method 1: Snapshot Retention

Snapshot retention refers to a retention method based on a snapshot of time of user activity on-site. This is the starting point for many and the most basic way to capture retention as a single metric.

DAU / MAU Ratio

The DAU / MAU ratio is an important industry benchmark and standard ‘snapshot’ retention metric. DAU & MAU are simply the number of unique users who engaged with your brand during a period of time.

DefinitionDAU (Daily Active Users): tracks the number of unique users seen in a 1 day window (usually yesterday)

DefinitionMAU (Monthly Active Users): tracks total unique users active in a single month (typically a rolling 30-day window to allow for insight not only by calendar month, but by a “windowed” 30-day period).

Once you have your DAU and MAU, retention is simply the ratio of DAU to MAU. Intuitively this ratio reflects the proportion of the monthly (or last 30 days) user-base that is active on a single day.

Example
A ratio of 1 means that every user who visited in the last month is active again today.

Depending on your industry or vertical, you can determine what a good benchmark DAU / MAU ratio is. For example, for a mobile app, 20% DAU / MAU is considered a good place to be. Over time you will develop a sense of what a good ratio for your business looks like.

Day-over-Day Retention

Day-over-Day (DoD) Retention is simply a more granular snapshot metric. With DoD, we simply take the ratio of users active today & yesterday to the total users active yesterday. This ratio represents the proportion of visitors that return to your site over a very short timeframe.

Example
Daily retention rate = # of users active today & yesterday / # users active yesterday

Limits of Snapshot Retention

The snapshot retention methods we’ve looked at are a great place to start and require less heavy lifting (assuming you have basic tracking in place). Because of their simplicity, certain shortcomings surface quickly:

  • All users active during the snapshot are included in the calculation regardless of the date of their first engagement with your brand.
  • Little to no distinction is made between existing habitual users and brand new visitors.
  • Short term and long term metrics can be greatly affected with little insight as to the cause.

Example:
If you have a one day marketing ‘burst’ campaign that brings a significant amount of new users, then your MAU metric will grow (which, on the surface is generally viewed as a good thing). It’s also likely that the ‘burst’ users will have higher churn than average. The net effect is the DAU / MAU metric will be greatly depressed. The worst part is the one day burst campaign will affect the retention metric for 30 days without any insight into why.

 

Method 2: Cohort Retention (the Classic Method)

Cohort retention avoids the pitfalls of snapshot retention by tracking the lifetime of groups of similar users – aka cohorts. Tracking the date when each user first visits your brand’s site enables apples-to-apples comparisons about the retention of the product.

While more involved, this method reveals deeper levels of understanding of the user retention-rate at specific times relative to the user start date (note the “start-date” could be the first time a user visits or other important milestones like registering for an account).

By re-analyzing users through the lens of cohort analysis, the surfacing of new insights, like these listed below, are now a possibility:

  • “Overall 50% of new sign-ups are still active (retained) 4 weeks after the signup date.”
  • “New user from the first week of January had an average retention of 4.5% at day 90.”
  • “Customers acquired during our goodyear-blimp marketing campaign were retained at an astonishing rate of 91% after one year.”

Calculating classic cohort retention

Typically, cohort retention is calculated as the ratio of users active at specific time period to the total number of users included in the cohort. This method is sometimes called ‘sample day retention’ or just ‘classic retention’.

Example:
Day 7 Retention = User active 7 Days after start-date / Total user active on start-date

A cohort’s retention curve represents the average lifetime so far of those people. You can get this super useful curve by looking up cohort retention-rate at multiple points (e.g. day 1, day 2, day 7, etc. -the more the better the ‘resolution’ of the curve)

Pro tip:
If you set a limit on the maximum user lifetime (i.e. retention rate = 0 at the maximum allowed lifetime), then ‘connect the dots’ with all the observed retention rates you arrive with a full life-time retention curve. The area under that curve is equal to the average lifetime for that cohort. (Checkout numeric integration here: http://mathworld.wolfram.com/NumericalIntegration.html)

NOTE: There are many variations of cohort retention that are worth being aware of:
 – Classic retention (the method focused within this post for being the most prominant)
– Full retention
– Rolling retention
– Return retention
– Bracket-Dependent Return retention

 

Method 3: Survival analysis

Okay, so the third method is the sole subject of another article. In a follow-up post we’ll introduce a ‘serious’ retention model – Survival Analysis. In the follow-up post we look at how your marketing analytics can benefit from a powerful tool developed for medical studies back-in-the 1960s – These techniques are especially relevant for dealing with the smaller data-sets typical of marketing tests.

email marketing CTAs

Shares
Share This