User testing tools aren’t just for user testing experts.
In fact, five-second tests offer a quick, easy and (relatively) cheap way for copywriters and digital marketers to run quick checks on their copy (“copy validation”).
Most of us know we need to use data to help us know what to write. But you should also use data to help you spot if you hit or missed the mark.
That’s where copy validation comes in. When you validate your copy, you boost confidence in your work knowing that it’s making a great first impression – long before your client ever reacts to it.
Time (as well as this study) says we have as little as one-tenth of a second to make a good impression.
Psychology Today and Business Insider say we have about 7 seconds.
We see a face. We form a first impression. It’s human instinct.
Unsurprisingly, that first impression is largely emotional. And a recent study by neuroscientists regarding how our brain forms a first impression confirmed through neuroimaging that the amygdala and the posterior cingulate cortex (PCC) are key regions of the brain involved in that process. And if you’re not someone who throws “amygdala” or “PCC” around in conversation, let me nutshell this for you: both do a lot of heavy lifting in processing emotions.
(Source: Kurniasanti, Kristiana Siste, et al. “Brain Regions,” licensed under CC BY-NC 4.0, Medical Journal of Indonesia, 2019.)
(Source: Photo by Robina Weermeijer on Unsplash)
And the amygdala? Well it’s hailed as the integrative center for emotions, emotional behavior and motivation.
But breaking down the emotions behind a first impression is a little more complex:
A recent study at Princeton University by Janine Willis and Alexander Todorov involved a series of experiments focusing on judgments formed during a first impression drawn from a face. Those judgments were:
- Attractiveness
- Likeability
- Competence
- Trustworthiness
- Aggressiveness
So participants looked at a picture like this:
And then judged it on those five criteria above.
Interestingly, only 1 of those 5 judgments actually has anything to do with facial appearance: attractiveness.
Willis and Todorov’s methodology was as follows:
- Participants were told that this was a study about first impressions and that they should make their decisions as quickly as possible
- The instructions explained that photographs would be shown for short periods of time and that the experimenters were interested in participants’ gut reactions
- Participants were asked a yes/no judgment question – for example “Is this person trustworthy?”
- Following this yes/no judgment, the next screen asked participants to rate their confidence level in their judgment
They found that increasing exposure time from 100 to 500 ms increased confidence in the participants’ judgments. But that increase in exposure did not change their initial judgment. In fact, they found that the judgments formed after just 100 ms of exposure corresponded with judgments made in the absence of any time limits.
Or put simply:
First impressions are sticky.
Lizard brain strikes again.
Of course, your first impression could be wrong.
As social psychologist Dr. Leslie Zebrowitz states:
“We seem unable to inhibit this tendency [to make quick judgments] even though it can lead to inaccurate impressions […] and has significant social consequences.”
We’re talking about consequences like:
Judicial decisions.
Financial success.
And even election outcomes
Similarly, a study conducted by Todorov and Jenny M. Porter found that even minor, random variations in images of the same person resulted in different inferred personality impressions from their participants.
So they saw faces like this:
And for each face, they identified if the person looked like a mayor, a consultant, a villain, etc.
(You can find actual photos, etc from the study here: Todorov, A., & Porter, J. (2014). Misleading First Impressions. Psychological Science, 25(7), 1404-1417.)
In his book, Todorov shares his insights:
“What appearance-influenced voters are doing is substituting a hard decision with an easy one. Finding out whether a politician is truly competent takes effort and time. Deciding whether a politician looks competent is an extremely easy task. Appearance-influenced voters are looking for the right information in the wrong place, because it is easy to do so.”
What’s happening here is simple:
We’re looking for the easy button.
And just like that wrong snap judgment you made about that guy with the resting bitch face (who now happens to be your best friend), you can just as easily form a wrong first impression of a website.
But unlike that friend, you probably won’t give that website a second chance.
Because in the case of your copy on a website’s home page, “wrong” usually means a visitor bouncing or (worse) a missed lead or sale.
Is our first impression of a website really that similar to our first impression of a person?
In short, yes.
A recent study researching the speed at which we form opinions about a web page’s visual appeal concluded that we will assess this within 50 ms.
And this study confirms that the judgments made during our first impression of a website do in fact influence our perceptions of credibility and trust.
Much like our ability to draw inferences from facial appearances, our first impression of a website is also comprised of a complex process that quickly accesses a variety of components. Many of them are design, but your copy also plays a part.
If we consider that most users leave a web page in the first 10 to 20 seconds, the implications become quite clear. As Jakob Nielsen puts it:
“To gain several minutes of user attention, you must clearly communicate your value proposition within 10 seconds.”
The necessity for a clear value prop in your hero section isn’t news.
How to ensure your copy is making the right first impression: copy validation
Usability analyst Craig Tomlin suggests that there are three pieces of critical information a new visitor should be able to answer:
- Who are you?
- What product/service do you provide?
- What’s in it for me?
Likewise, Peep Laja of ConversionXL suggests that your first impression should communicate:
- Where the visitor is
- What they can do there
- Why they should do it
- Your brand personality (chic, silly, sexy, savvy, smart, classic, etc.), and
- Your differentiating factor (what makes you different from the competition)
So, where does the five second test fit into all of this?
In his book The UX Five-Second Rules, Paul Doncaster says the advantages of the five-second testing method include:
- Speed
- Efficiency
- Portability
- Flexibility
- Simplicity
- Convenience
(Those advantages were music to my ears the first time I read them.)
And while Doncaster notes that the original reason for this test was to simply confirm that the purpose of a content page was obvious, this testing form can also be used for bigger, more critical website components.
Like the hero shot of your homepage.
And he notes that, because we’re testing an “in the moment” response, you don’t need to have access to a bunch of existing customers with prior knowledge of the product or service you’re selling. It’s all about visual perception and short-term memory.
So, given what we know about the seemingly uncontrollable snap judgments our brain makes on our behalf that influence our behaviors and decisions, the five second test simply provides an opportunity to receive feedback on the information a viewer is gathering during those first critical moments.
And I repeat: you don’t need to be a user testing expert to start running these tests.
In a nutshell, the five second test is a type of survey methodology – you’re going to be asking questions and gathering responses.
That’s not so hard, right?
It’s basically like a really teeny tiny version of a customer interview. But with anonymous strangers. And copy that’s in development.
But here’s the thing:
Just because you’re sending your copy out into the wild for validation does not mean it has to be “ready” or “perfect.”
(Heck. It doesn’t even have to be done.)
This is about innovation. And it’s about trying to solve problems (hopefully without creating new ones).
So pull out that long list of value prop options, and choose a few standouts to test. Because we’re going to use an iterative design process to quickly gather feedback and continue writing.
The goal here is to improve through iteration… not just document all of its flaws.
So don’t get hung up on perfection.
(And, yes. I realize that is so much easier said than done. This is why you’re going to see some of my less-than-perfect tests below. Hi, my name is Carolyn and I’m getting comfortable being vulnerable.)
(If Kate Winslet can admit it, I can too.)
But, as with other testing methods, this test has limitations.
“When all you have is a hammer, everything starts to look like a nail.”
Five second tests aren’t always the perfect solution.
Similar to the way we form a first impression when we see a human face, in a five second span we are able to take in a lot of information… but may not necessarily be able to make sense of it. As Doncaster writes:
“A participant may take in a lot of information perceptually, but likely does not have time to make much sense of it as a whole entity, resulting in feedback that is limited in scope.”
In understanding the limits of a five-second test, you’ll be able to set up a test that effectively supports your learning goals.
A five second test can’t tell you everything.
And though this is perhaps obvious, it’s not the right tool to choose when the question you’re trying to answer requires more than five seconds of thought or consideration.
Hint: This means that body copy is not a good fit for this method of testing.
Where the five second test fits in your R&D tool belt
Conversion copywriters, this is where it gets particular important for you.
Five second tests are a great tool to add to your research-and-discovery tool belt for the simple fact that they allow you to focus on the first touchpoints between your copy and your audience.
And, as data analyst Tomi Mester notes in the article linked above, it’s worth mentioning that you’re not looking for statistically significant results here. You’re looking for insights that can help you form an educated guess. And steer you away from bad ideas dressed in sheep’s clothing.
And while Doncaster notes that homepages were originally “off limits” for this testing method, I would argue that because we’re testing an “in the moment” response, homepage hero shots and value props are ideally suited to this type of testing. Because this portion of your website is so critical to a visitor’s first impression of a website – and by extension of the company. Not to mention the success of the overall website as a whole.
So the five second test becomes an opportunity to:
- Settle an internal bar where you have a team or internal stakeholders that have strong opinions about which approach they think is best
- Gather intel about how clearly your copy is communicating an aspect of your desired message
- Check the memorability of said message
- Check the clarity of said message
- And help guide your selection between value prop versions to begin validating before your copy goes live (or even graces your client’s desk)
Sure. You could wait to run A/B tests on your live copy (if traffic and conversion rate allow).
But why start with an experiment, with all its costs and risks, when you can first validate? Once you’ve validated a message, THEN split-test it.
The five second test offers immediate feedback and a chance to ditch poor performers – and test variations that may tank – before they cost you leads or sales.
How To Set Up Your Five Second Test So You Know Your Copy Is More Likely to Work
Here’s the step-by-step approach that I, a conversion copywriter who relies heavily on data and has run dozens of these tests, use when running 5-second tests:
- Select copy for testing
- Define testing goals
- Select the right 5 second test format
- Write the test
- Run the test
- Process the results
- Determine next steps
- Rinse and repeat (as necessary)
Now assuming you’re interested in running tests like these, here’s a breakdown of what to do at each step.
1. Select your copy for testing
For the purposes of ensuring a good first impression, you’ll want to choose copy placed at first touchpoints between you and your reader.
Me? I use this method primarily for testing homepage value prop copy.
While first impressions come from more than just headline copy, it’s already been noted here on Copyhackers that your value proposition is second only in importance to the visitor’s motivation for landing on your site in the first place.
With this in mind, it makes sense to test the clarity of your value prop’s first impression within the context of the hero shot.
As Joanna says:
“Visitors to your site need to be told what’s unique or different about you that they’d really like.”
With our gauge on likeability beginning to form during those first few milliseconds, ensuring the clarity of your homepage copy messaging (which is likely to be a version of your value prop) becomes all the more critical.
Because here’s the thing:
Ensuring that your value prop is making a good first impression will improve your chances for success (i.e., reducing bounce).
Running a series of five second tests helped me write this homepage hero copy:
The results? A 63.51% drop in page bounces.
Remember: good first impression = improved chances for success.
Other “First Impression” Copy Elements You Could Try Testing
- Landing page headlines
- Sales page headlines
- Any other headline you might be working on
- Facebook ads
- Instagram ads
- Adwords ads
- Email subject lines
2. Define your testing goals
In the wise words of Lewis Carroll:
“If you don’t know where you are going, any road will get you there.”
Similarly, if you don’t clearly define why you’re testing and what you’re hoping to learn, it’s unlikely that you’ll gather the results you’re looking for.
(More importantly, you risk wasting time and money.)
In the context of making a good first impression, you’ll most likely be defining your testing goals as one of the following:
- Is my value prop memorable?
- Is my value prop clear?
Tip: If you’re stuck on defining a testing goal, look to the 5 critical points of a USP and headline scorecard. Your low-scoring criteria becomes low-hanging fruit for testing.
3. Select the right five second test format
Based on your defined testing goal, you’ll select a test type.
Doncaster outlines the following four test formats:
- Memory dump test (i.e., what is most remembered overall)
- Target identification test (i.e., what is most remembered about a specific visual target)
- Attitudinal test (i.e., perceived appeal, quality and/or usefulness)
- “Mixed” test (i.e., combining aspects of the previously mentioned types of tests)
Could your five second test succeed without defining a test type? Maybe.
But in doing so, you largely leave the results of your test up to chance.
Defining your test format not only helps you more clearly align your learning goals with your results, but it also helps you write a better test.
You don’t need to overthink this. Here are some guidelines to help you choose your test format:
- Want to gain insights about the memorability of your value prop? Choose memory dump.
- Want to gain insights about the clarity of your value prop? Choose memory dump or target ID.
- Want to gain insights about the emotional response you’re eliciting? Choose attitudinal.
You’ll notice I didn’t mention the “mixed” test. Here’s why:
As Doncaster notes, it’s simply more difficult to create a solid “mixed” test and, by default, more difficult to gather useful insights.
Not impossible. But more difficult.
So in the interest of saving budget and working to eliminate non-responses, your best bet is always to stay within one test format. With that in mind, your rule of thumb should be:
One goal = one test format
4. Write your test
Just so you know, this is where most five second tests fail.
There are two critical written components to your test (aside from the copy you’ll be testing):
- Your welcome screen
- Your test questions
Doncaster’s book sets forth his testing rules based on the analysis of over 300 public online five second tests. Of the eight “violations” listed, five of them were concerned with the writing of the test.
Here are some best practices to keep in mind as you write:
1. Your welcome screen: keep it simple
This is the screen that your test participant will see immediately before viewing your test image. It’s at this point that you’re setting the stage for what’s to follow.
While this is certainly an important moment in your participant’s journey through your test, you can (and should) keep this simple.
Just like your value prop, your instructions should be clear, concise and specific. Here are some examples pulled from Doncaster’s rules, organized by test format:
Sample Memory Dump Test Welcome Screen:
“You will have 5 s to view the [image]. After, you’ll be asked a few short questions.”
“You’ll see a screen for 5 s. Try to remember as much as you can about what you see.”
Or, the basic UsabilityHub.com welcome screen message:
“Look at the interface for 5 seconds and remember as much as you can.”
Important: At the time of writing, UsabilityHub.com (my chosen five second test tool) limits custom welcome screens to researchers with a paid monthly membership.
Sample Attitudinal Test Welcome Screen:
“You will have 5 s to view the [image]. After, you’ll be asked a few short questions about your reaction to the [image/design/message].”
“You’ll see a screen for 5 s. Pay attention to the general appeal of the [image/design/message].”
“After viewing the [image/design/message] for 5 s, you’ll be asked about your opinions on its look and feel.”
Sample Target Identification Test Welcome Screen:
If you are planning to only ask questions about one element among an image with many elements:
“You will have 5 s to view the [image]. After, you’ll be asked a few short questions about the [specific target being assessed].”
That said, if you’re hoping to understand whether the specific target is easily identifiable within the full design:
“You’ll see a screen for 5 s. Pay close attention to the [image/design/message].”
“After viewing the [image/design/message] for 5 s, you’ll be asked about your opinions on its look and feel.”
A word of warning:
Be careful that the information provided on the welcome screen doesn’t prime your users’ responses.
For example, if your welcome screen reads:
“Imagine you’re assessing companies that provide house cleaning services. You’ll see a screen for 5 s. Try to remember as much as you can about what you see.”
And then your first question reads:
Q1. What service does this company provide?
Well, you’ve primed the users response by tipping them off about what they’re about to see in the instructions. In turn, this will bias your results and waste the moment when your user’s memory is most clear. So don’t do that.
And on the topic of writing questions:
2. When it comes to your test questions, less is more
First things first: there’s no magic number for how many questions you ask.
That said, in the case of five-second tests, less is more.
The trace decay theory helps to explain the memory fade taking place in your participant’s short-term memory bank. Simply put, unless the information being transmitted is rehearsed, the short-term memory can only hold that information for 15 to 30 seconds before it begins to decay and fade away.
With this in mind. it’s best to ask the least number of questions possible to answer your research question and satisfy your learning goal. The original test questions used by Perfetti included only two questions:
- Recall as much as you can remember about the design.
- What is the purpose of the page?
In his book, Doncaster notes that of the 300+ tests he analyzed, 80% asked three or more questions. While most five second test tools cap out at five questions, know that you don’t need to max out your question capacity.
Doncaster makes a strong case for the “less is more” approach:
“In five-second tests, the acts of reading, understanding, and responding to questions place additional burden on the cognitive process in play, which contribute further to memory fade.”
Only ask what needs to be asked to satisfy your learning goals.
Lesson Learned: It’s not just about the number of questions, it’s also about their order
The final questions on your five second tests are like the (questionable) leftovers in the back of your fridge: A little fuzzy and likely to elicit an “I dunno” response. 🤷🏻♀️
We’re going to look at one of my very early five second tests.
My goal was to improve the clarity of communicating Consulting by Hart‘s services.
One of the issues with the existing copy that we identified during the initial project scope was that visitors couldn’t clearly identify the intended audience. As in, when it came to answering the question “Am I in the right place?” visitors may not confidently answer “yes.” The five second test run on the control hero shot helped us confirm this.
Here’s the control hero shot:
We started by asking them this question: “What do you remember about the site?”
And here’s a look at the results to Q1:
My goal was to shine a spotlight on the services provided.
My original question order for this series of tests was:
Q1. What do you remember about the site?
Q2. What can you do on this site?
Q3. Who is this site for?
Q4. What is the name of their business?
Notice any issues here?
For starters, Q4 doesn’t tie to the testing goal.
I also learned, through the gathered responses, that Q3 was open to misinterpretation. Some of the respondents interpreted this particular question as meaning “Who is this site representing?” or “Who is the company that this site is for?”
But the biggest issue?
Some of the most specific questions were asked at the end of the test. While the end results gathered were still promising (60% of the participants identified the name of their business correctly), this did leave my test with an increased vulnerability to participant non-responses.
Your solution: Ask the questions requiring the most specific memory recall first.
Your participants’ memories will be most clear immediately after your testing image disappears. From there, your participants’ short-term memories will begin to fade.
Doncaster dubbed this the “reverse-Polaroid effect” and noted that writing with this effect in mind helps you combat the “I don’t know” or “I don’t remember” responses.
If I were to do-over this test today I would rewrite those questions to look something like this:
Q1. What is the service or product this company provides?
Q2. What else do you remember about the image?
Here’s why:
Because the goal was to improve the clarity of the offer, I’ve removed all of the questions that do not specifically tie back to that goal.
And because the test is structured in a memory dump format, you’ll notice that I’ve now re-ordered the remaining questions according to specificity. This allows me to capitalize on the test taker’s memory when it’s strongest.
Finally, I’ve tweaked the wording of these questions to make my ask a little clearer. And adding the “else” in Q2 allows the participant to dump anything else they felt was memorable.
TL;DR: Think carefully about the questions you ask and the order in which you ask them. Order questions from most specific to least specific.
5. Run the test
You’re done the pre-work. Now it’s time to unleash your test on the public!
Here’s the step-by-step method to follow:
1. Load-in your test
If you haven’t already done so, you’ll need to choose your five second testing tool.
I like UsabilityHub.com because it allows you to pay a small fee for platform-recruited test-takers. Meaning you don’t have to bug your friends, family, mom or colleagues to take your test.
(It’s free to bring your own participants to your UsabilityHub test, but it still doesn’t open up the option to write your own instruction screen.)
If you’re bringing your own participants, you may want to consider UserBob. It allows you to set up a test with a custom instruction screen and then pushes your test taker toward your chosen survey tool (like Typeform).
Lesson Learned: Choose your visual context carefully
You have a couple of choices when it comes to presenting your copy to your test takers:
- Test copy on a blank screen
- Test copy in a wireframe
- Test copy by editing copy on the existing site using an editor extension
(We’re going back to one of my first wireframed tests. I’m sweating.)
As previously mentioned, my goal on this project was to improve the clarity of the services my client provided.
Here’s that wireframed copy:
And here are some of the lacklustre answers I received to the first question:
Highlights include:
“The first thing was the big giant “X” going through the page.”
“It’s shaped like an envelope.”
Not exactly the great-first-impression-confirming insights I was hoping for. Womp, womp.
It wasn’t that the answers as a whole were completely unusable, but it became immediately apparent that my very lo-fi wireframe spoiled the results.
The solution: Test your copy in a visual context that isn’t going to distract your participants.
Here’s the thing:
Those spoiled results were entirely my fault. And I knew it as soon as they landed in my inbox. So I went back to the drawing board to try again. This time I tested my copy in a wireframe that actually resembled a website:
And the resulting answers were decidedly more focused:
Unsurprisingly, improving the visual context for the copy improved the overall quality of the answers I received from participants.
TL;DR: Choose your visual context with care. I like running tests on my copy in context because that’s how a new visitor will actually experience it. If you decide to run your wireframes, make sure they look as much like a website as possible.
2. Set your audience demographics
This step is only required if your testing tool will recruit participants on your behalf.
Your gut will tell you to narrow down the participant demographics as much as possible. The common thinking here is that you’ll receive more qualified feedback if you set your demographic filters to match your target audience.
Stop. Do not pass go. Do not collect $100.
Yes, validating your copy with your market is important.
But this isn’t the time for that.
This is about first impressions. Is the message clear? Is the message memorable?
That’s it.
You don’t need your testing participants to be carbon copies of your one reader. You just need them living in the same (figurative) neighborhood. Why? Because most demographic data doesn’t help you understand your typical visitor’s behavior.
Within your testing dashboard you’re likely to see at least some of the following options to filter your audience:
- Language
- Viewing device
- Country
- Gender
- Age range
- Education level
- Employment status
- Annual household income
- Technical proficiency
- Daily hours online
Of these options, I focus first on behavioral indicators, like education level, employment status, technical proficiency and daily hours online.
I also set the viewing device filter. This helps ensure that the participant is viewing my image in the correct visual context – i.e. I don’t get participants viewing a desktop hero shot image on a mobile device.
Depending on the project, I might also set the language settings to English. I’ve found that this can help protect my test results from increased “I don’t know” or “I don’t remember” responses from participants that might have English as a second (or third) language.
Bottom line: Skip the granular “female, aged 25-34” details and focus on big-picture behavior-driving demographics instead.
3. Set your audience size
This was one of the biggest questions I had when I first started using this testing method in my writing process.
The jury is out on the “perfect” audience size for testing:
- Virzi says 3-5
- Perfetti says 20
- Faulkner says 10
- Turner et al. says 7
- And Nielsen says 5
After reviewing the research, I think that Nielsen’s suggestion of five is the sweet spot for our purposes.
Here’s why:
Nielsen suggests that your first five test participants will discover over 80% of all problems.
I won’t get into the specifics of Nielsen and Landauer’s formula (you can find that formula as well as the their graph demonstrating diminishing returns as new participants are added here), but usability experts do seem to agree that simple tests require fewer test takers.
Sticking with five test takers also helps you make the most of your testing budget. As Nielsen states:
“Doesn’t matter whether you test websites, intranets, PC applications, or mobile apps. With 5 users, you almost always get close to user testing’s maximum benefit-cost ratio.”
Granted, this decision is based on usability studies, not studies specifically relating to copy. But the fact remains that our use of the five-second test qualifies as “simple,” so using Nielsen’s rule of thumb works.
And, as Nielsen also suggests, from a cost-benefit perspective, you’re better to run plenty of small iterative tests than one large test.
I’ve experimented with larger audience sizes, but keep coming back to five participants.
Here’s why five is the magic number:
During my work with Portica, a project management tool built specifically for architects and designers, its founder mentioned how he felt they were missing the mark in effectively communicating how they differed from more widely adopted document storage solutions. They weren’t really a document storage solution, but their existing messaging didn’t clearly articulate the full scope of what it is they actually offered.
I decided to double my typical testing audience size on this particular project because we were limited to only a small handful of customer interviews. That meant that my VoC data was drawn largely from raw mining research and founder interviews. I felt that a slightly larger audience pool might help confirm whether or not I was on the right track with their new messaging.
Here’s round one, version one:
And here’s round one, version two:
What I found as I toggled in and out of the audience filters on my completed results surprised me.
The filtered responses of smaller response pools remained consistent with the full audience results.
4. Preview your test
This little step can save you from wasting time and money on a faulty test.
Here’s a quick checklist to help you run quality assurance (QA) on your test:
- Welcome Screen
- Are your welcome screen instructions clear, concise and specific?
- Have you unintentionally primed your test taker with leading instructions that you should revise / remove to minimize bias?
- Testing Image
- Is your testing image fully visible without scrolling?
- Does your testing image display your copy in a visual context that accurately represents the way in which a reader would experience it? For example, if you’re testing homepage hero copy, does your testing image resemble a hero shot??
- Is there anything in your testing image that has the potential to distract your test taker?
- Test Questions
- Are your questions clear, concise and specific?
- Are your questions in the right order for your learning goals?
- Have you unintentionally primed your test taker with leading questions?
- Are there any questions that are unnecessary in helping you satisfy your learning goal? (If so, remove them.)
This step takes about a minute total to complete:
But that minute can save you from wasting budget on a problematic test. Don’t skip it.
5. Let your test run
Arguably the easiest part of the process: hit submit and wait.
If you’ve opted to have your testing tool recruit participants on your behalf, you can typically expect your complete results in the next 5 to 45 minutes.
Go grab a coffee, your work here is done (for now).
6. Process your five-second test results
Your results are in! It’s time to get back to work.
There are a couple of different ways you can go about processing the results.
While the easiest and quickest way is to review your results inside your test dashboard, I prefer to process results in a Google sheet (you can find that template here) in order to have a full picture of my control and USP iterations as I work through my projects.
Inside that worksheet I note down my pre-work decisions, test conditions and any demographic filters placed. Once I receive my results, I begin making my notes and tallying repetitions in language used by the test users (similar to Ashley Greene’s method demonstrated during this Tutorial Tuesday).
If you’re anything like me, I’m sure you’re wondering:
How will I know if my results are “useful”?
I’m a firm believer that you can learn from all of your results.
Receiving non-responses like “I don’t know” or “I can’t remember” is a signal to head back to the drawing board and try again.
Likewise, if you receive a set of responses that indicate you’ve achieved your goal, that also becomes useful (and handy to report back to your client).
For example, watershed protection and preservation organization Friends of the Muskoka Watershed came to me asking for snappy copy that articulated their rather complex mission and methodology in a way that is easy to understand.
I first confirmed their hunch about general audience confusion during the homepage hero control five-second test (as well as audience interviews and reviewing traffic flow through Google Analytics).
I then found, through various research activities, that their “science to action” methodology was particularly attractive to their target audience and was also a key differentiator between them and their competitors.
From there I worked up a variety of customer-facing options and began testing.
Here were the wireframe images I started with:
First the “snappy” version:
Followed by the longer version:
What I found?
Only 20% of the “snappy” version test takers understood their core mission – protecting the Muskoka watershed.
Compare that to 60% of the longer version test takers.
These results supported my recommendation for longer copy. And I continue to develop and test this copy as we move toward a finalized project.
7. Determine your next steps
Did your shift in messaging help you achieve your learning goals?
- Yes? Great. Move forward feeling more confident in knowing that your copy is communicating what it needs to communicate in those first five seconds of a visitor landing on the site.
- No? Also great! Head back to the drawing board with that intel to continue working on your value prop.
Use those findings to inform future iterations of your copy.
8. Rinse and repeat those five-second tests often (and when necessary)
In the wise words of Eugene Schwartz:
“Copy is not written. Copy is assembled.”
An engineer wouldn’t release a new product without testing it, right? You are an engineer of words. Testing should be built into the process of assembling.
As Nielsen says:
“After creating the new design, you need to test again. Even though I said that the redesign should “fix” the problems found in the first study, the truth is that you think that the new design overcomes the problems. But since nobody can design the perfect user interface, there is no guarantee that the new design does in fact fix the problems.
A second test will discover whether the fixes worked or whether they didn’t. Also, in introducing a new design, there is always the risk of introducing a new usability problem, even if the old one did get fixed.”
Everywhere he says “design,” replace it with copy.
I’m feeling a little like a broken record at this point, but here goes:
Your five-second tests are experiments.
Test early. Test often. And don’t let that perfection monster rear his ugly head!
Test your copy before you think it’s perfect.
And maybe even before you think it’s “ready.”
I typically start testing two or three customer-facing value prop options during my initial research and discovery work (before I present my messaging recommendations report), simply because it allows me to present some initial results to help guide my client in understanding why my recommendations suggest that one direction might be better than the other.
And it allows me to validate what I might think of a good idea as unclear and improves my work along the way to finalized copy.
The end result?
Copy that is communicating key information clearly within those first critical moments of a visitor landing on a website.
The critical deciding factors in whether or not you should test should be:
- Do you know what you want to learn?
- Will displaying an image, with copy, for 5 seconds help you learn this?
If you answered yes to both of those questions, then you should start testing.
This all sounds great. But how do I get budget approval?
Easy. I build my five-second test budget into the fee for my project.
This is now a non-negotiable in my writing process. So I don’t ask for permission.
Of course, exact numbers depend entirely on the project in question, but with each test costing about $10 USD, I usually build a $50 to $70 internal “Five Second Test Fund” into each project.
But if you want to dig in and get started right now, remember this:
“The best results come from testing no more than 5 users and running as many small tests as you can afford.” – Jacob Nielsen
User testing tools aren’t just for user testing experts. To get started with improving your copy’s first impression, you just need to be ready to learn and have some copy to test.
It’s all part of the iterative process on your way to making that great first impression.
Author: Carolyn Beaudoin, conversion copywriter
Editor: Joanna Wiebe, founder of Copyhackers
Peer reviewer: Talia Wolf, founder of GetUplift
Conversation