This guide walks you through using New Relic to optimize the performance of those digital properties that represent the end of your customer journey. It's part of our series on observability maturity.
This is the second guide in our Customer experience practice. For the first guide, see Quality foundation.
Bottom-of-the-funnel analysis is about improving conversion by focusing on performance at the end of the user journey.
Most sites and apps are built with one or more purposes in mind. If there's a purpose, there's likely to be a user journey. Some examples:
|Download a whitepaper
|Raise a support ticket
|Customer services (registration, forms, payments)
|Make a payment
|Stream a movie
|Buy clothes and accessories
|Find a support program in your state
|Join distribution list to get promotions and coupons
|Get information about events in your town
|Services (travel, rentals, bookings)
|Book a flight
|Share a selfie
When a user completes a journey, we think of it as a conversion. All conversions have a value - from a few dollars to thousands. The best way to improve the conversion rate is to start at the bottom of the funnel, when the intent to complete the user journey is clear.
What if I have performance issues in the middle of the funnel?
Any glaring issues with your app or site should be addressed, no matter where they occur. When it comes to optimizing conversion, it's better to start at the bottom of the funnel for two reasons:
- Higher return on investment. Users at this stage are already more likely to convert. Addressing performance issues here will see an immediate impact on the bottom line.
- If you have issues at the bottom of the funnel, optimizing earlier stages may not have much of an impact on the conversion rate.
Once you've optimized the bottom of the funnel you can use the same techniques to optimize earlier stages of the user journey.
Optimizing the top or the middle of the funnel without focusing on the bottom first is a bit like fishing with a holey net. You can get more fish into the net by optimizing when and where you fish but you risk losing all that optimization as soon as you take the net out of the water.
Increase revenue by resolving issues that appear when a user attempts to complete an action.
Key performance indicators
Bottom-of-the-funnel analysis measures the following KPIs:
Bottom-of-the-funnel success/conversion rate
The rate of conversion once a user has gone far enough to demonstrate intent to complete an action to actual completion. Examples of this are rates of:
Increase the rate of conversion by addressing errors and latency at the bottom of the funnel
Revenue at risk due to latency
Value of a conversion times the number of pages or interactions in the bottom of the funnel that are slower than the industry threshold.
Focus on reducing this value by improving page KPIs
Revenue at risk due to errors
Value of a conversion times the number of backend errors in the bottom of the funnel interactions
Tune this value to make it meaningful by filtering out errors that aren't visible to the end user. Once this is meaningful, focus on reducing it. Create an alert to notify you if it suddenly trends upward.
You should have:
- Familiarity with New Relic's synthetic monitoring
- Familiarity with our
- Familiarity with basic New Relic browser monitoring UI views
- Familiarity with New Relic's SPA data in the browser monitoring UI
Required installation and configuration
You should have:
- Browser agent (Pro version) installed on relevant pages
- SPA enabled for single page applications
- Data retention for browser events greater than or equal to 2x an average sprint
Establish current state
The steps you'll do to establish the current state rae:
- Identify where the bottom of the funnel starts
- Distinguish between pages and actions
- Create a scripted synthetics monitor for the bottom of the funnel
- Import the bottom-of-the-funnel dashboard
- Capture current performance
These steps are described in more detail below.
Identify where the bottom of the funnel starts
The bottom of the funnel is focused on the final steps of a user journey where a user has gone far enough to show intent to complete the journey.
Here are some examples.
Ecommerce user journey
The user journey is simplified so you can focus on where the bottom of the funnel begins: at checkout. Most users entering the checkout phase plan to purchase something. Reducing errors and latency from this point onward is more likely to improve conversions than focusing on any other part of the funnel.
Car insurance purchase user journey
In the example above, you have the user's interest in car insurance as they enter information, but you don't know their intent until they see the quote and continue to proceed.
Distinguish between pages and actions
The final steps of a user's journey is likely to be a mix of full page loads and AJAX calls.
You'll need to know all pages and AJAX requests for the next step. If you're not sure which requests are running from the page in question, you can run:
SELECT count(*) FROM AjaxRequest WHERE pageUrl like '%FILTER%' FACET groupedRequestUrl SINCE 1 DAY AGO
Create a scripted monitor for the bottom of the funnel
Make sure you have a scripted monitor for each path through the bottom of the funnel. The goal is to make sure your bottom-of-the-funnel services are working around the clock.
For example, you may have a checkout flow that calls a different payment API depending on the customer's payment preferences.
Import the bottom-of-the-funnel dashboard
Follow the instructions in the bottom-of-the-funnel analysis quickstart to install the dashboard and configure it.
Capture current performance
- Follow the instructions in our bottom-of-the-funnel analysis README on GitHub.
- Use the dashboard from the previous step to understand the bottom-of-the-funnel performance.
- Create a plan to improve KPIs that don't meet target values, as well as reduce revenue at risk.
Plan your work
Whether you have a dedicated initiative to improve performance or classifying as ongoing maintenance, you need to track your progress at the end of every sprint. For detailed information, see:
A common question is: Should I apply segmentation?
Segmentation (breaking out performance into cohorts, such as region and device type) is a good idea if:
- Your organization has initiatives tied to addressing a target audience from a particular cohort that you can segment by using either custom attributes or data that is already available in New Relic.
- You are already familiar with bottom-of-the-funnel analysis and there's a significant enough difference in performance among different cohorts to warrant tracking and/or developer focus.
Best practices going forward
- Revisit performance metrics at the end of each sprint.
- Any time the user journey changes, revisit if the bottom-of-the-funnel steps are the same. Incorporate changes in developer sprints as needed.
- Openly share metrics with the lines of the business you support as well as other internal stakeholders.
- Revisit your customer experience SLOs. Do you need to define more ambitious SLOs for the end of the funnel?
- Create for business critical drops in Quality foundation KPIs.
At the end of this process you should:
- Know your user conversion rate and have addressed errors or performance issues that negatively impact it.
- Increased revenue for your company.
- Created, or be on the path to create, a common language with business owners so you're working together, which will open new avenues for recognition and sponsorship for new projects.