Imagine an entire focus group of diverse users who never sleep, work for free, and deliver feedback instantaneously.
Sounds great, right?
With synthetic personas (also known as synthetic users), we are closer to this possibility than ever before. This means faster and more flexible testing, validation, and iteration. And for your business’s bottom line, this translates into more efficient conversions, more leads, and more revenue.
But before we get carried away, what actually are these synthetic personas?
Synthetic personas explained
The basics
Synthetic personas are based on an understanding of real users. In an ideal world, they will match exactly with your existing personas, faithfully recreating their motivations, behavioural patterns, and interests. Creation of a synthetic persona will begin with the identification of this persona, a process anyone in UX design is already well familiar with. Pre-existing data on this persona is then collected from any available sources. These may include:
- Google Analytics data
- Previous user testing sessions
- Interviews with existing users and audience members
- Recordings of user journeys taken from Microsoft Clarity or similar tools
Some more advanced synthetic personas may even be given additional data, such as psychological OCEAN profiles, or relationships with other personas within the same team!
Data matters
The word “synthetic” is not accidental. These users are assembled by aggregating hundreds of data points on real user journeys, categorised by persona, motivation, and intent. This dataset provides your AI tool of choice with immediate access to real evidence of how their fictional synthetic persona will behave in each particular scenario.
To use all of this data, synthetic personas engage a process known as retrieval-augmented generation (RAG). This essentially means that an AI agent is referred directly to a predefined set of information before it begins generating a response. This might include user data, company information, or product details. This ensures that the response is not just a guess of what a given persona might say but rather an actual attempt to simulate that persona based on real data.
Fast, cheap, agile
The benefit of investing into an accurate set of synthetic users is that they will then provide a source of immediate, flexible, and on-demand testing for any pages or components being designed. This means that design teams can quickly troubleshoot designs during the design process, allowing for remediations to be made far faster and cheaper than otherwise possible. The cheap unit cost of synthetic personas also means that they can test components that are too small or insignificant to warrant expenditure on “real” user testing.
How synthetic users can improve your conversion rates and lead generation
More testing, more information
This should go without saying, but the more information you can gain about your conversion funnel, the better positioned you’ll be to improve it. And a more efficient conversion pipeline translates directly into more leads and more revenue. The conventional method for collecting this information is A/B testing, but this takes time, risks providing real users with sub-standard experiences, and can only provide insights into one feature at a time. Synthetic personas, on the other hand, deliver instantly valuable information without affecting your existing conversion pipeline. Plus, this information is presented contextually, saving you the struggle of interpreting why one option outperformed another.
Learning from success
Synthetic personas are very, very useful, but we shouldn’t forget that they are not actually our users. As the saying goes “the map is not the territory”. The usefulness of these representations relies on their connection to real, accurate, user data, and as soon as that connection begins to shift, the synthetic personas will lose their value.
It is crucial, therefore, to continually retrain your synthetic personas on fresh, updated, behavioural data. The best way of doing this (although it is a bit complicated), is to systematically measure the similarity between synthetic and real responses to the same tasks and interviews. This then allows you to carefully adjust your synthetic personas, increasing their parity with actual human responses.
Synthetic personas will completely transform what it means to test, experiment and validate digital solutions. The question is not if, but when. As of yet, these personas do not offer a complete replacement for old-fashioned validation methods, but are already a great option for quickly testing and prototyping designs.
If you want to build a more efficient conversion pathway, and believe synthetic personas can help, just get in touch with our UX team today. We’d love to chat!
