Measure marketing funnel success

Step-by-step Playbooks
, , , , ,
1

Based on How to Create a Marketing Funnel by Responding to Customer Behavior by Tom Whatley

Business benefits

Consistently improve and optimize your marketing funnel with high quality data.

Calculate cost per acquisition (CPA) for paid marketing and email marketing using this formula: CPA = (cost of your campaign) / (number of conversions)

Calculate your churn rate using this formula: (number of customers lost) / (number of customers at the start of the period)

Calculate average revenue per user (ARPU) using this formula: (total revenue from all users) / (number of users)

For example, if you have 100 subscribers, 50 of whom are on a $50 level and 50 of whom are on a $100 level, your ARPU is $75 a year.

Calculate lifetime value using this formula: LTV = ARPU x customer lifetime

Measure conversions for each stage of your marketing funnel.

  • TOFU: The number of visitors that convert to marketing qualified leads (MQLs).
  • MOFU: The number of MQLs that sign up or become subscribers.
  • BOFU: The number of sign-ups and subscribers that become customers.

Measure the success of each channel in your marketing funnel, ensuring that every channel has a clear conversion goal.

For example, with SEO, clicking on a link could be classed as a conversion. With email, the conversion might be registered when a recipient responds or clicks through to your website.

Track user behavior with heatmaps and screen recordings to learn how people are using your site.

Heatmaps show how people move around and interact with your website. Using them can help you identify which parts of your landing pages work and where changes need to be made to better optimize for conversions.

Use heatmaps to measure:

  • Engagement: How visitors interact with your blog posts and landing pages.
  • Actions: Which buttons and links visitors click.
  • Attention: Which headlines, images, and forms get the most attention.

Use screen recording software such as Hotjar or LogRocket to see how real users engage with your website or app.

Look for patterns that harm the customer experience. Obstacles like broken images, slow loading pages, or confusing or misleading copy may cause frustration that causes a person to leave without converting.

Combine all of the data you gathered to get a complete picture of user behavior. Identify areas of the marketing funnel that need attention.

Last edited by @hesh_fekry 2023-11-14T15:15:43Z

2

Given that we’ve just calculated churn rate, shouldn’t this use LTV = ARPU / churn rate instead? As per How to Calculate Customer Lifetime Value - The LTV Formula - Baremetrics ?

3

In the article you posted it does state that you

can also calculate lifetime value using churn (which is a number you likely have more readily available).

LTV = ARPU / User Churn

There are several ways. They also suggest using SAAS tools like Baremetrics. We use a combination of Metorik and ChartMogul to calculate this as well as things Like customer churn rate etc…

4

It gets a bit lost in the playbook, but in the blog post, calculating the churn rate is presented as part of calculating LTV:

Measure the retention value of your customers by looking at:

Churn rate: The number of customers that stop paying in a given period (e.g., If you had 200 subscribers and lost 10 in the last year, your churn rate is 5%).

Average revenue per user (ARPU): The average revenue of all current accounts (e.g., If you have 100 subscribers, 50 of whom are on a $50 level and 50 of whom are on a $100 level, your APRU is $75 a year).

Lifetime value is then calculated as follows:

ARPU x customer lifetime = LTV

Churn rate seems more a distraction than a useful calculation at this point.

I can see how having a customer-loyalty KPI would be useful, but the blog post isn’t recommending that. So it’s basically saying, Calculate churn rate. Now do absolutely nothing with it.

Edited to add: We also haven’t told them to look up or calculate the average customer lifetime.

5

True. It seems a bit of an outlier. Unless we are both missing something here. Good question to ask for sure.

Hopefully can get Tom Whately to clarify.