Adlib: The latest technique in my UX research toolbox

Exploring an underused and lesser known research technique

This research technique was something I first came across when attending a UX Research conference hosted by General Assembly and presented by Michele Ronson back in 2020. It is only now, 2 years on, that I have started to use it, and I really do wish I had tried it sooner.

Mad libs, Ad lib, complete the sentence, fill in the blank. Call it what you like (I will continue with Mad libs as that is what it was called in the conference), however the last two are a lot clearer around the concept of what it is.

It is, in essence, a pre-constructed sentence or paragraph with key words missing throughout. It is up to the participant to fill in the blanks with their views or opinions or experience.

I have included an example of how a mad lib can be put together just to give you a flavour of its potential.

INTRO: You will be required to fill in the blanks in the paragraph below with your own personal experience and responses. Please answer honestly and if there is anything you are not comfortable with sharing just write ’N/A’

Hello my name is _______. I am _____ years old. I currently work as _____ which I have been doing for ______ (please specify time frame eg months, years). The main challenges of my jobs are _______, _______, ______ (list your top 3 challenges). Despite them, what keeps my keeping coming is _______, _______, _______ (3 things you enjoy at work). In an average week I am in the office for ______ (days)

From my point of view it has three key strengths:

1. It is a way of offering a a stronger sense of guidance in an unmoderated test by giving the participants a discussion guide of sorts. They have the script we would normally steer the conversation with, but in its purist form.
2. It is also a great way to get participants warmed up and into the testing mindset before getting them to interact with a prototype, moving them into the topic and experiences they will have had. As this is something some of us did when we were at school, it helps create a more relaxed and fun environment.
3. It is a more personal approach to getting questions from participants as they are invested in the end to end script, as they have complete visibility of it and how all the answers are connected. The way it is structured is made to feel like a very personal story about them and so the answers they provide are made to feel more important

Rules of creating a Madlib

1. Don’t just jump into the question you want to know. It’s important to remember, this isn’t a survey. It is not a Q&A format and so you need to slowly ease the participant into the question you want to know by progressively moving the conversation that way.
2. Set boundaries where applicable. This is primarly done by following up the blanks with the boundaries in brackets and italics eg (please specify your answer in days). This makes it easier for the participant to know what you want, and easier for you to analyse as it removes some of the clutter from their answers. One note of caution though, when you do this, make sure you are not priming the user with an expected answer in the brackets.
3. You can follow up on their questions by including them on the same page as the madlib. In the above example if I wanted to explore the ‘why’ to the section what keeps me keeping coming is _______, then underneath the madlib I could write Please explain your answers to ‘what keeps you coming back’.
4. Make sure that the script flows. Similar to the first point, it is important to remember this isn’t just a standard survey question. It is a story of a user and so it’s important that it flows like one. Make sure there is a flow from one section to the next, or chronologically if it’s exploring a user journey.

Learns

I have used this method a couple of times in a remote unmoderated environment, and each time the results were really easy to read through and compare as well as simple to set up. I also found that, compared to other methods of remote testing, there were fewer ambiguous or irrelevant responses because it was a lot clearer for participants to know what information we required.

That said you need to make sure that you really define what sort of response you need to make sure it is easy to analyse afterwards. For example specifying the time frame or asking the participants to specify the time frame they used so you know what they meant.

You should also not expect long detailed responses, rather shorter answers, which is why exploratory questions can be used to follow up underneath.

Having done this a couple of times now, I would say a recommended amount of blank spaces would be in the region of 20 (to prevent fatigue bias). This value, however, is purely based on my 2 tests so there is flexibility on that value based on your circumstances and requirements.

THE PERSON BEHIND THE WORDS

Its always encouraging to see your work recognised, so if you found this article interesting, useful or just a general good read, please do leave a clap or follow. I hope to write more content soon and this will be the best way to find them!