Skip to main content

We are currently experiencing an issue that is preventing customers from submitting support tickets. Please contact us at (866) 819-4486 if you need immediate assistance.

x
Optimizely Knowledge Base

Improve metrics that matter with your optimization program

This article will help you:
  • Align your optimization program with goals that are important for your business
  • Quickly build a goal tree using our templates, based on industry verticals
  • Decide where to focus your optimization efforts

When you build an optimization strategy for your company, you'll first align your program’s testing efforts to your company goals. Without a strategic guide, you run the risk of building unfocused or inefficient tests that consume time and resources without producing results. How do you ensure that your tests will make a real difference? Where should you focus your efforts?

We recommend that you start by focusing your testing program on achieving quick wins with impactful tests that affect the bottom line. A goal tree can help you meet this challenge.

A goal tree is a map of the metrics that are critical to your company’s success. The discovery process for creating a goal tree helps you translate business metrics into concrete optimization goals.

Each goal breaks down into a set of smaller goals. For example, "Revenue" in the tree above cascades into "Revenue per Visitor" and "Number of Visitors." Top-level metrics roll directly into smaller KPIs that you'll optimize for.

You can track revenue through an Optimizely goal! Check out our article on revenue tracking.

A goal tree helps you:

  • decide where to focus your optimization efforts
  • break down broad organizational imperatives into bite-sized tests
  • find multiple avenues to generate impact through optimization
  • optimize for smaller KPIs and understand how these impact your business goals
  • help stakeholders understand the value of your optimization efforts in a language they understand

You only need to build a goal tree once, but you’ll consult it multiple times throughout the iterative optimization cycle. You’ll use it to judge the impact of each potential campaign or experiment when prioritizing your backlog or building your roadmap. You might consult it when designing specific tests. You’ll probably also examine it when analyzing your results to consider the impact on key performance indicators (KPIs). Use the goals in your tree to identify other departments that care about the same metrics; these departments may be able to offer additional resources and support for your program. You can also use the tree to communicate optimization goals and gains to an executive sponsor.

The optimization program manager should be responsible for updating the goal tree if your company’s business model changes. Team members who focus on A/B testing strategy and analysis will also work closely with this document.

What you need to get started:
Materials to prepare
    • Company-wide and business unit key performance indicators
    • Benchmark performance metrics: conversion rates, RPV, LTV, AOV
    • Leading indicator metrics: test velocity, efficiency, quality, agility
    • Program budget
People and resources
    • Program manager
    • Team leads (technical, design, analytics)
    • Executive sponsorship
Actions you'll perform 
    • Benchmark KPIs and baseline modeing
    • Goal trees exercise
    • Align with stakeholders on optimization program goals
Deliverables
    • Team charter with program, team, and individual goals
    • Revenue tree or similar outline of hierarchy of goals
    • ROI Models
What to watch out for
    • Many organizations have an internal debate about the most important metrics, but it's important to focus on the right top-line goals
    • Leading indicators are important to test, but often poorly understood and under-tested

Building a goal tree

Ready to create your own goal tree? Download a goal creation workbook. Choose the template for your industry vertical:

Choose the model that best matches your business. To modify the template, ask yourself:

  • How does my company generate revenue?
  • How does each metric break down into a smaller metric?

Remember that each category of the tree branches into smaller categories that should equal the original.

Using a goal tree

We recommend that you focus your testing efforts on the KPIs at the bottom of the tree. These metrics - like average order quantity - are opportunities to optimize for profitability. Moreover, focusing on smaller pieces rather than the top metric can help you raise the win rate, since experiments are faster to run and more likely to reach statistical significance. The lift you generate rolls directly to the metric at the top.

 

It’s a good idea to combine your goal tree research with data from your analytics platform. This can help you figure out where to optimize for the biggest impact. Imagine, for example, that the KPI you’re targeting is average order quantity.

Maybe your analytics tell you that certain products are purchased in larger quantities: non-perishable pet food versus expensive televisions, for instance. You may want to test a strategy that promotes those products, by offering a discount on pet food bundles or suggesting certain purchase quantities on the checkout page.

If you see improvement that confirms your hypothesis, consider leveraging this insight through personalization. On your landing page, display a banner that suggests that visitors will build up a winter stock of pet food; show this banner to visitors who've browsed pet food in the past. Match your site experience to SEM campaigns that promote high-value pet food brands. Create humorous pet food campaigns for younger audiences, and value-oriented messaging for discount shoppers.

Goal trees focus your optimization strategy on concrete metrics, so your team’s efforts to design and manage campaigns and experiments make an impact.

 
Note:

Looking for more resources on identifying the right company metrics and understanding your funnel? Check out this workbook or this Optiverse post.