Published on November 21, 2017 and updated on October 19, 2021.
Traffic volume on a website or app isn’t enough to tell how many people will perform any valuable action. This metric is usually paired with others in a funnel that represent the steps from a journey that end on a sale, register, meeting or any other valuable action for a business. This model presumes that traffic must run across each of the funnel steps to reach the desirable outcome. The proportion of the traffic that cross each step is called conversion rate, a metric that is at the center of a CRO project. CRO means Conversion Rate Optimization, and define a set of techniques, tools and methodologies that are systematically used to optimize conversions. The term was coined in 2007 by the Conversion Rate Experts.
The main premise which guide the work of CRO professionals is that there is always something to improve. A CRO project (or program, as some agencies may call) have an end, but leaves for legacy a culture of constant improvement. Optimize conversions may be seen as a work without end, since there is always something to improve. Each CRO project must have its own definition of end, so the final result can be compared with the one obtained before the start of the project in order to evaluate its overall success.
Apart from that premise, other aspects are equally important for a successful CRO project:
- Systematic procedures are the rule: the end result must be verifiable, comparable and replicable, which is only possible with a process that is well defined and followed;
- Each business is a business: the same procedure will produce results that are aligned with each business’s reality, which are different by nature;
- Question the numbers: quantitative data show only what happened, not the reason. Use other tools to gather qualitative data that can improve the results of the experiments.
Experiment management in CRO
Concepts like MVP, rapid experimentation and lean are all based on the ideas of learning fast and following a process with well defined steps. These processes take on the idea of cientific experiments applied on business settings. CRO projects also drink from that source. The end result is a backlog of experiments to be tested against hypothesis, not unlike what is seen on other areas, like Growth Hacking.
There are many tools that can help manage CRO projects. Some of them are:
- Pipefy: Provide a framework for experiment management that can be used inside their app. The steps are: brainstorm, prioritize, prototype, test, learn;
- Moz: Provide a similar framework. The steps are: data collection, hypothesis generation, prototype, implement and test, analyze;
- Conversion XL: They built a framework designed for CRO projects, focused on data quality. The steps are: heuristic analysis, technical analysis, analytics, mouse tracking, qualitative research, user testing.
No single framework is the definitive, and it is possible to combine steps from each of those frameworks to support different business needs.
Tools for CRO projects
There are many options from tools, ranging from cheaper to expensive and from friendly to technical. This opens up the possibility of testing different tools in search of the stack that will better suit each project’s needs. Bad project executions can generate noise in the data collected, which can invalidate any findings. There is a lot of content online to help setup a good “laboratory” for studies. The challenge then is to know which tool to use for each scenario.
The CRO stack
There is a great post on Medium that list 32 tools for CRO, placing each one within a stack, with 4 sets: quantitative tools, qualitative tools, testing tools and workflow tools. A CRO project will need at least one tool from each of the sets. I’ll list some of the tools which I’ve worked with, but I recommend reading through the Medium post to find out many more tools.
Quantitative tools
The quantitative tools stack is comprised of tools that collect and store behavioral data for a product. They are used to understand behavioral patterns. Some of these tools are:
Qualitative tools
Quantitative tools show what people are doing with a product. Qualitative tools, on the other hand, try to explain why those behaviours occur. Some of these tools are:
Testing tools
A running experiment is a test to challenge one hypothesis. Testing tools are the necessary components to run hypothesis tests. Some of these tools are:
Workflow tools
These tools help track the CRO project, as well as the status of each of the experiments. Some of these tools are:
CRO on organizations of different sizes
CRO projects rely on information about the business and its target customers. This data, along with knowledge passed from the business, serve as basis to formulate hypotheses to be verified. Even small businesses have some data to work with in order to generate some hypotheses. Different conversion optimization strategies will be adopted according to the size of the business. Entrepreneur Brian Balfour wrote a blog post in its website explaining how different testing and learning strategies evolve based on the size of the business.
Learning strategies based on experiments proposed by Brian Balfour for each business maturity stage. Source: Brian Balfour.
He suggests that tests, as a whole, should be big, specially at companies that are at the beginning. Examples are tests with user flows, different customer segments and value propositions. Once the company knows its business model, as well as its customer base, tests can be more segmented, based on a more narrowed growth strategy.
CRO on apps
CRO projects on apps aren’t that common, although nowadays have some popularity. This happens for at least four reasons:
- Code inside apps is compiled, making its internal structure hard to access using testing tools designed for web applications.
- It is necessary to be approved by all main app stores and follow their app publishing policies, making the process more bureaucratic.
- User need to download / update the app to see the changes.
- Data may not be readily available, since users can turn off the internet and use apps offline.
Even with those drawbacks, there are tools to work specifically with mobile tests, including some of the tools that were listed before. Two of these tools are:
The optimization part in CRO
Optimize in CRO goes beyond the metric that defines conversion. Optimization also happens, on my point of view, when the CRO project leave as a legacy the need and discipline to always learn more the right way, that is, producing knowledge that can be shared, verified, trusted, replicated and useful. The key metric that represents conversion is a deliverable, but if the organization is not capable of sustaining the optimization strategy — or change it when needed — the effort spent on a CRO project will not suffice, and it wont be better than doing tests without well defined procedures at all.
Additional resources
- ConversionXL have a very resourceful blog about CRO;
- MOZ is another company that provides an extensive list of resources on the subject;
- Qualroo created a CRO course;
- Rock Content published an e-book with CRO fundamentals.
Let’s optimize!