The origins and application of the FACT & ACT model

A common pitfall in conversion optimization is a lack of structure. Where do you start? What is your ultimate goal? And how do you ensure that the knowledge gained does not get lost in the process? Especially as your CRO program grows and more people become involved in the process, a lack of structure can greatly reduce the quality of your optimization work.

From the beginning at Online Dialogue, we have run hundreds of AB tests, reaching millions of people. Based on all our insights and experiences, we have developed a framework for AB testing: FACT & ACT. The model structures daily CRO practices and focuses on working with a multidisciplinary team. In this article, I explain how Online Dialogue's FACT & ACT model came about, how it has developed over the years, how we apply it ourselves and how we teach it to our clients.

The scientific method

Scientific research is usually empirical in nature and aimed at testing hypotheses. A hypothesis is formulated which may or may not be rejected by observations after which, if necessary, a new, modified hypothesis is formulated. Again, this hypothesis will then be tested based on observations, and so on.

Models are also used in science to structure scientific research. One of the most commonly used models is the Empirical Cycle which describes the different stages of the research process in the empirical sciences. ‘Observation - induction (hypothesis formation) - deduction (prediction) - testing - evaluation.’

Since in CRO we also work on confirmation through empirical observations (evidence-based growth), we took the scientific model (see the image above) as the basis for our own FACT & ACT model.

The development of the FACT & ACT model

The first version of the FACT and ACT model was a linear model. But soon we found out that this did not work the same way in practice. If you test a lot, you learn a lot and you can use the new information to speed up your process. And so came up with the round FACT & ACT model. The inner circles of the model rotate faster than the outer circles. So you gather more and more information faster and faster, allowing you to make better decisions and thus also optimize the process.

FACT & ACT step by step

How do we use the FACT & ACT model? You always start with a thorough data analysis. Here you collect all the data (qualitative and quantitative) that you have. About this data you have assumptions, which we call main hypotheses (if you do [this] then [this] will happen, for [this] reason). You will then test these main hypotheses.

If you have a good main hypothesis, you want to quickly find out if your assumption is correct. So you start testing. In our experience, there are so many confounders (when a third factor, related to both the determinant and the outcome, disrupts the causal relationship between the two) that you want to test one component more often. One test is not a test.

Find: Define the customer journey and its potential. What and where will you test?

Analyze: Analyze behavior using research and find hypotheses.

Create: Create designs for testing based on the hypotheses.

Test: Test your designs.

Analyze: Analyze your results based on the test results.

Combine: Combine your results with the insights about hypotheses.

Transform: Transform your product/service/communication with your new insights.

Then you can start drawing conclusions. The last, but certainly one of the most important steps (often forgotten, by the way), is to share the results with the rest of the organization. The information gained from experiments can be applied at all levels of the company. In addition, the new insights will be the input for new tests.

From Transactional to Transformational

For a good CRO program, it is important to have a clear goal. When companies are just starting out or have not been doing Conversion Optimization for very long, the initial goal is really always to generate money. Over time, as the maturity of the business grows and you've already done a lot of testing, the knowledge gained will also generate value in other ways (for example, in the form of a new product or service). In short, if you test a lot, you learn a lot and you can do more with that knowledge than just rake in money.

Transactional

At the first level, the focus of your CRO program is on achieving the business case. We therefore call this the transactional phase. By following the model in this way, you can easily scale up the number of tests to achieve maximum growth.

This transactional phase can yield a lot of gains, but you're missing the focus on learnings. The many tests you run will certainly yield lots of loose insights, but you need to do more to filter out general learnings from your results. This is where the strategic model of the FACT & ACT model is valuable. In this phase, you combine more data sources and insights than just your Web analytics. Think for example of scientific articles, UX research, old learnings, feedback from visitors and so on. For this we use the 5V model to run through all data sources in a structured way.

Behavioral Intelligence Graph

The ultimate goal is a ‘Behavioral Intelligence Graph’: a model that plots your visitors' behavioral drivers and influencers for each step in the customer journey. Based on this model, you formulate hypotheses focused on earnings, learnings and new insights. In this step, it is also important that you start prioritizing your hypotheses (see prioritization models like: PXL, PIE and ICE) and testing based on the evidence you have already gathered from previous research and/or testing. You continue to experiment and the results continue to feed your Behavioral Intelligence Graph. As a result, the model grows more and more into a solid behavioral model that explains what drives customers to use and value your product or service. With these insights, the CRO team (or even the entire company) can formulate better and better hypotheses.

Transformational

In addition to improving your CRO program, such a behavioral model provides the information you need to begin transformations. First within the company, by providing valuable information for marketing purposes, technological developments and product innovations. But also in the market itself, you are able to share insights that change how companies offer their services. That's why we call this Customer Intelligence-focused level ‘the transformational level. You transform the business with the customer insights you've gained. And so you have a data-driven way of working, developing, improving and transforming. No longer based on gut feeling, but based on data.

The FACT & ACT model has already proven its worth with many of our clients and also in our trainings it has taken a central place. Want to see how the model is applied in practice? Then read the Hostelworld Case Study.