Experimentation / CRO

Experimentation and Testing Programs acknowledge that the future is uncertain. These programs focus on getting better data to product and marketing teams to make better decisions.

Research & Strategy

We believe that research is an integral part of experimentation. Our research projects aim to identify optimization opportunities by uncovering what really matters to your website users and customers.

Data and Analytics

90% of the analytics setups we’ve seen are critically flawed. Our data analytics audit services give you the confidence to make better decisions with data you can trust.

The Presentation of Your Value Proposition Matters [Original Research]

Many people in the marketing space are trying to figure out how to best present their value proposition. Which copy works best? Which design?

These are important questions because your value proposition is such a high impact area of your site - some would say the most important part. So even though many people are working on and researching value proposition presentation, we thought there was still room to investigate.

This study manipulates the value proposition of a financial service SaaS website, and uses eye-tracking and survey tools to test differences and effectiveness among the value prop. variations.

Background

Basically, a value proposition is the cream of the crop when it comes to selling your product.

It's the facets and features of your product that set you apart from the competition, that define your product, your brand.

A value proposition can boost your sales considerably when done right. A value proposition is a promise of value to be delivered. It’s the primary reason a prospect should buy from you.

Although the concept of a value proposition is easily grasped, the best way to present a value proposition is still quite debatable. There's a need for data on users' perception and consumption of value propositions presented in different formats.

What's the best way to present your value proposition?

Data Collection Methods and Operations

We applied two methods in this study:

1. Eye-tracking: to understand if simply changing what is included in the value proposition changes the way people process the page, in what order and for how long they pay attention to the value proposition.


Numbers of participants with usable eye tracking data included in analyses
Numbers of participants with usable eye tracking data included in analyses

2. Questionnaire: After viewing a value proposition, we asked participants about their opinions on the design, which included questions to testing memory recall rates of the site's benefits and features.

We manipulated the landing page of personalcapital.com to create 4 different value proposition designs (see them below). Designs were placed in the same area and all other aspects of the landing page remained untouched to maintain consistency among treatments. When beginning the study, participants were prompted with these task instructions:

"Hello!  You will be shown a company web page. Browse as if you were considering the company's services. Afterwards, we'll ask you a few questions."

Design Variations (screenshots | red boxes are to highlight the relevant differences, they weren't visible or included in the experiment):

Variation A: 3 Bullets
headline + 3 bullets with desc
Variation B: 3 Bulleted Descriptions
5 Bullets
Variation C: 5 Bullets
Description Paragraph
Variation D: Only the Short Paragraph

  1. After viewing this page, what is the purpose of this website? (open-ended response)
  2. What is you confidence level in knowing who this service is for? (7 point Likert scale)
  3. What can you tell about the services this company offers? (open-ended response)
  4. What is your confidence level in knowing what the features of the service are? (7 point Likert scale)
  5. Did you notice any particular advantages to using this company's services? (open-ended response)
  6. What is your confidence in knowing what the benefits to you are? (7 point Likert scale)
  • Questions 1 and 2 were created to identify how well the value proposition communicates the product's purpose.
  • Questions 3 and 4 were created to identify how well the value proposition communicates the services included in the product.
  • Questions 5 and 6 were created to identify how well the value proposition communicates the unique features of the company's product.

7.       Participants were then shown all 4 variations and asked to select the design they most preferred. Normally, we ask recall in this situation, but a preference test works in this case because we're ultimately interested in how useful customers feel about the information presentation.

Findings

Eye-Tracking results revealed users' gaze path among the major areas of interest (generally indicated with the different color boxes) as they browsed the web page:

Eye-Tracking:

Variations:

3bullet variation
Variation A: 3 Bullets


Capture
Variation B: 3 Bulleted Descriptions


5 bullet variation
Variation C: 5 Bullets


paragraph variation
Variation D: Only the Short Paragraph

In the figures above, the numbers represent the order in which people, on average, viewed the different elements. Notice the order changes among the variations. The trend is that the shorter the text block of the value proposition, the longer it takes to get noticed.This isn't surprising at all, considering that we expect elements that take up more interface real estate to attract more attention faster.

Let's look at the numbers. Our primary area of interest here is the 'key message,' where the value proposition (the site's benefits and services) is explained in varying levels of detail.

  • Average Time to First Fixation (first notice) – varied significantly among treatments [F(3,190) = 8.8966, p < 0.001]. It took participants an exceptionally long time to notice the value proposition in the variation with only the short paragraph.
Average time to first fixation and standard deviation, per variation
Average time to first fixation and standard deviation, per variation

Takeaway – user’s first time to notice the value proposition varied based on the content and layout of the proposition. The more text in the value prop., the quicker user’s noticed.

  • Average Total Time Fixating (how long users paid attention to the message) – varied significantly among the treatments [F(3,190) = 17.0814, p < 0.001].
Average total time fixating and standard deviation, per variation
Average total time fixating and standard deviation, per variation

Takeaway – user’s total time reading the value proposition varied based on the amount of copy, content, and layout of the proposition. The more text, the longer the user read.

In other words:

reading stats on key message

Variations in total time ‘reading’ don’t necessarily indicate that participants absorbed more information from one variation over another. To identify that, we need to explore the survey results:

Survey Questions:

  • Likert Scale Questions

Questions 1-3 yielded inconclusive results, meaning that there were no significant differences between variations among viewers' mean confidence levels for knowing:

1. Who the service is for

2. What the features of the service are

3. What the benefits of the service are

Takeaway - users did not differ in their perceived confidence in understanding the value proposition.However, the open-ended responses provided different insights!To test for differences among the responses for the variations, we took the answers of the following questions and coded them as such:

Open-ended Question 1:

After viewing this page, what is the purpose of this website?

binary (yes/no) coding: 0 = did not understand at minimum the site was about providing financial planning tools, 1 = did understand this. Used a Chi-squared Goodness of fit to test for differences in 'task completion' among the variations.

Open-ended Question 2:

What can you tell about the services this company offers?

Interval (think 1,2,3) coding: 0 = no services listed, 1 = 1 service listed, 2 = 2 services listed, etc. Used a Kruskal Wallis test (non-parametric version of ANOVA) to test differences.

Open-ended Question 3:

Did you notice any particular advantages of using this company's services? Interval coding: 0 = no advantages listed, 1 = 1 advantage listed, 2 = 2 advantage listed, etc. Used a Kruskal Wallis test to test differences.

Open-ended Question 1 Finding:

"After viewing this page, what is the purpose of this website?"Here is a sampling of some of the actual responses and how we coded them:

Understood the site's purpose:

"To be able to work on your financial goals with everything in one place"
"Financial monitoring"
"To manage my personal financial accounts with their online tools and calculators all in one place."
"The website provides free financial tools and calculators so someone can reach and view their goals and all financial aspects."
"A tool to help keep you organized with your finances and financial investments"
"Tools for financial success"
"To get a complete view of your personal finances"
"Selling personal financial services. Offering a dashboard of your overall financial health"

Misunderstood or didn't know the site's purpose:

"No idea"
"I'm not really sure what the purpose of the website is for."
"Buying electronics"
"Organization on laptops and technology?"
"Help with electronics?"
"This to me looks like a banking page website."
"For connecting pc computer"
"Office work, or supplies"

A Chi-squared test found differences among the variations in users' average understanding of what was the purpose of the website [F(3,215) = 14.666, p = 0.002]. Here are the numbers for 'pass'/fail' (really 'understand'/'didn't understand').

Numbers of users who passed (understood) and failed (misunderstood) what the purpose of the website was for after viewing
Numbers of users who passed (understood) and failed (misunderstood) what the purpose of the website was for after viewing

Takeaway - The more information provided in the value proposition, the more people were able to understand the purpose of the website/service.

Open-ended Question 2 Finding:

"What can you tell about the services this company offers?"

The Kruskal-Wallis test indicated that there were differences in the average number of services recalled among the value proposition variations [H(3) = 15.374, p = 0.0015].

Mean number of services recalled for the different value proposition variations.
Mean number of services recalled for the different value proposition variations.

We see that the two winners here are the '5-bullet' and '3-bullet with description' variations, which are the variations with the most information/services listed. These results also correlate directly with the amount of time fixating from the eye tracking results.

Takeaway - People are able to recall more services when there is more information regarding services to read.

Open-ended Question 3:

"Did you notice any particular advantages to using this company's services?"

The Kruskal-Wallis test indicated that there were differences in the average number of advantages described among value proposition variations.

Mean number of advantages described by users among the variations
Mean number of advantages described by users among the variations

Again, value propositions with more information ('5 bullets' & '3 bullet with description') resulted in more thorough survey responses.

Takeaway - People are able to describe more website advantages when there are many features and benefits to read about.

Preference Question:

"Which website variation do you prefer"

A Chi-squared goodness-of-fit test indicates that there were differences in user's preferences depending on which original variation they saw [X2 (9, N=215) = 24.231, p = 0.004].

Numbers of preferred variations according to the original seen. Red boxes indicate highest preference count per variation.
Numbers of preferred variations according to the original seen. Red boxes indicate highest preference count per variation.

What's interesting here is that people who viewed the '5 bullet' variation preferred the '3 bullet with description variation and vice versa.

Overall, people tended to prefer the simple bullet variations (without description). A binomial test with confidence interval indicated both at near significance at a 95% confidence level for greater than the expected proportion of preference (25%).

The 5 bullet variation had 29% of votes (Confidence Interval = 24.6% - 34.8%; p = 0.066) and the 3 bullet had 28.5% of votes (Confidence Interval = 23.7 - 33.8%; p = 0.118)

Takeaway - variation preference depended on which variation was originally was seen, although users generally prefer bullets; both versions with only bullets received high preference counts.

Limitations

There are copious testable variations on value proposition design, we only tested four of them. Perhaps other formats communicate a company's services far more effectively. It might be interesting to study the effects of dynamic value propositions (propositions that employ moving graphics, gifs, or videos).

Additionally, there may be questions we didn't ask that could have uncovered further results.

Identifying participant background (employment, age, etc.) might reveal patterns among understanding of this particular site's value. Since we only tested one site and one product, our findings may be due to the particularity of the website at hand. For example, if we tested a common, widely-known product (cosmetics or office supplies), it's possible that results would've yielded a better understanding of the site's services or benefits.

Conclusion

How you present your value proposition matters.

In general, we found that the more descriptive the value proposition was, the quicker people noticed it and the longer they read it. They also tended to recall the information better, and they could describe more advantages of the website when presented with more information. When it comes to preference, however, the variations with bullet points and less information tended to win out.

The main takeaway: If you want people to actually read the value propositions you're composing, limit other elements on the web page. Elements that stay on the page should be extremely relevant to the value proposition's message.

Related Posts

Who's currently reading The Experimental Revolution?