Professional Documents
Culture Documents
Good UI Data Story Issue 6
Good UI Data Story Issue 6
Overview
Three tests were run across 3 months with the Control and 7 Variations
A H
+8%
Revenue
Best Variation
Measuring Generated $6.54 per visitor
Revenue
Control
Generated $6.08 per visitor
idea
51 Price Illusions
Removed dollar signs.
Possible Key
Difference
Exposing Fields
idea
20
Exposed the checkout fields on the first page instead of creating a separate checkout page.
Hence two pages were merged into a single one.
idea
35 Urgency
Button label states the goal and adds a sense of urgency (“Become a gold member now”).
idea
18 Benefit Button
Button was named as “Become a Gold Member” to carry a benefit.
Test Snapshot
Test 1 Test 2 Test 3
A
The control was
carried over across B
three tests
C
D
Variations
idea
49
Curiosity
idea
24
Selling
Benefits
idea
22
Gradual
Engagement
P2: Checkout
idea
50
Reassurances
idea
22 Gradual Engagement
Gradual engagement is used with the first page asking to make a more subtle choice.
idea
50 Reassurances
Site is reassuring security on the checkout page with the headline.
idea
49 Curiosity
The headline suggests a hint of curiosity.
B The B Variation & Its Effects
Variation B was a small change in the direction of keeping more focus. The fourth purchase option was
removed to see if it would drive more attention to the three subscription options.
Results are grey if they
aren’t conclusive (p-Value is
higher than 0.20). A -1%
The Control
idea
16
Keeping
Focus
Compared To A -15%
Revenue
idea
7
Recommending
idea
18
The Control
Benefit Button
idea
16
Keeping
Focus
idea
18 Benefit Button
Button was named as “Become a Gold Member” to carry a benefit.
idea
16 Keeping Focus
The footer with an additional purchase option was removed.
D The D Variation & Its Effects
Variation D variation was very similar to C and existed to check if recommending the first 1-Month plan
would make a difference.
Compared To A -32%
Revenue
idea
7
Recommending
Default Set
To 1Month
idea
18
Benefit Button The Control
idea
16
Keeping
Focus
idea
18 Benefit Button
Button was named as “Become a Gold Member” to carry a benefit.
idea
16 Keeping Focus
The footer with an additional purchase option was removed.
Compared To A -17%
Revenue
idea
7
Recommending
Default Set
To 3Months
idea
18
Benefit Button The Control
idea
16
Keeping
Focus
idea
18 Benefit Button
Button was named as “Become a Gold Member” to carry a benefit.
idea
16 Keeping Focus
The footer with an additional purchase option was removed.
idea
7
Recommending
The Control
Default Set
To 3Months
idea
18
Benefit Button
idea
16
Keeping
Focus
P2: Checkout
idea
24
Selling
Benefits
Easier Plan
Switching The Control
idea
18 Benefit Button
Button was named as “Become a Gold Member” to carry a benefit.
idea
16 Keeping Focus
The footer with an additional purchase option was removed.
idea
20 Selling Benefits
Three benefits were repeated at the top of the page.
Compared To A -17%
Revenue
idea
7
Recommending The Control
More Visible
Plans
idea
18
Benefit Button
idea
16
Keeping
Focus
idea
18 Benefit Button
Button was named as “Become a Gold Member” to carry a benefit.
idea
16 Keeping Focus
The footer with an additional purchase option was removed.
Be
The H Variation & Its Effects
st
Pe
rfo
rm
er
The main idea behind variation H was to take both pages and merge them into a single one while at the
same time allowing them to make an informed decision about each of the plans.
Results are green if
they are conclusive
Compared To A +8% (p-Value is higher than
0.05).
Revenue
idea
20 idea
Exposing 35
Fields
Urgency
idea
18 The Control
Benefit Button
Shifted
Alternative
Choices
idea
51 Price Illusions
Removed dollar signs.
Possible Key
Difference
Exposing Fields
idea
20
Exposed the checkout fields on the first page instead of creating a separate checkout page.
Hence two pages were merged into a single one.
idea
35 Urgency
Button label states the goal and adds a sense of urgency (“Become a gold member now”).
idea
18 Benefit Button
Button was named as “Become a Gold Member” to carry a benefit.
Hypothesis
Going into this test, we knew the traffic would be low and planned just 1 variation. The client also
suggested removing the “message pack” footer, so we added that as an additional variation. We started
with the idea that we could reduce the 3 buttons of choices to just 1 primary CTA. The other choices would
be secondary “options, which we would turn into radio buttons. This would allow our preferred option to be
pre-selected. We learned from the pre-tests that 1 month was the main choice with 90%+ of people choos-
ing that option. However, our assumption was that plans of longer duration held more value for the busi-
ness and should be the focus of our efforts. For example, we assumed monthly subscribers would quit
after 1-2 months. However, we did not have an opportunity verify this assumption. By making the 3-month
plan the default option, we hoped to pull some people from the 1 month plan to the longer 3 month plan.
Test Setup
A Converted at 15% (14% to 17%) with 271 out of 1765 visits p-Value Improvement
B Converted at 15% (13% to 17%) with 271 out of 1772 visits 1 -0.4% (-18% to 17%)
C Converted at 12% (10% to 14%) with 215 out of 1792 visits 0.007 -22% (-38% to -4.9%)
Data is shown at a 95% confidence interval and adjusted for multiple variations using the Bonferroni correction.
A Converted at 2.1% (1.4% to 3%) with 37 out of 1765 visits p-Value Improvement
B Converted at 2.1% (1.4% to 3%) with 37 out of 1772 isits 1 -0.4% (-50% to 49%)
C Converted at 3.3% (2.5% to 4.4%) with 59 out of 1792 visits 0.05 +57% (-2.3% to 109%)
Data is shown at a 95% confidence interval and adjusted for multiple variations using the Bonferroni correction.
A Converted at 0.7% (0.4% to 1.4%) with 13 out of 1765 visits p-Value Improvement
B Converted at 0.6% (0.3% to 1.2%) with 11 out of 1772 isits 0.86 -16% (-91% to 64%)
C Converted at 0.3% (0.1% to 0.8%) with 6 out of 1792 visits 0.17 -55% (-116% to -24%)
Data is shown at a 95% confidence interval and adjusted for multiple variations using the Bonferroni correction.
FINDING Footer Removal With 4th Purchase Choice Didn’t Have Any Effect
Removing the footer did not produce any notable effect on the revenue. Both variations A and B had a very
similar revenue per visitor. We did however still recommend removing the footer as it seemed to generate
more clicks to the three plans which was an indication of more people starting off in the sales funnel.
For this experiment, we clarified the prices and savings on Page 1, to create a smoother transition into
Page 2. We did this on all variations. Next, we wanted to know if simply changing the default would make
a difference. So, we set up a “first is default” and a separate “middle is default” variation.
We also surmised that the radio button is too subtle, that people didn’t pay attention to it, continued to the
default, and realized it’s not for them. So, in variation G we tried a compromise visual treatment, similar to
the Control in that it presented 3 distinct options but still with a preselected default, more concise text,
and still as a secondary choice.
In variation F, we also wanted to address the idea that people were going back and changing their plan
choice. We set up a variation identical to our “middle default” option on the first page, but added a way to
switch plans directly on the second page, so there would be no more going back.
Test Setup
A Converted at 20% (17% to 22%) with 215 out of 1089 visits p-Value Improvement
D Converted at 8.5% (7% to 10%) with 93 out of 1096 visits 0.0001 -57% (-71% to -42%)
E Converted at 14% (12% to 16%) with 156 out of 1121 visits 0.0003 -30% (-45% to -14%)
F Converted at 16% (14% to 18%) with 184 out of 1138 visits 0.028 -18% (-34% to -1.9%)
G Converted at 15% (13% to 18%) with 174 out of 1123 visits 0.009 -22% (-37% to -5.3%)
Data is shown at a 95% confidence interval and adjusted for multiple variations using the Bonferroni correction.
A Converted at 2.1% (1.4% to 3.2%) with 23 out of 1089 visits p-Value Improvement
D Converted at 7.2% (5.8% to 8.9%) with 79 out of 1096 visits 0.0001 +241% (145% to 301%)
E Converted at 4.3% (3.2% to 5.6%) with 48 out of 1121 visits 0.004 +103% (29% to 160%)
F Converted at 3.4% (2.5% to 4.7%) with 39 out of 1138 visits 0.059 +62% (-4% to 118%)
G Converted at 2.7% (1.9% to 3.8%) with 30 out of 1123 visits 0.39 +26% (-33% to 82%)
Data is shown at a 95% confidence interval and adjusted for multiple variations using the Bonferroni correction.
A Converted at 0.3% (0.1% to 0.8%) with 3 out of 1089 visits p-Value Improvement
D Converted at 0.3% (0.1% to 0.8%) with 3 out of 1096 visits 0.92 -0.64% (-125% to 124%)
E Converted at 0.5% (0.2% to 1.2%) with 6 out of 1121 visits 0.36 +94% (-83% to 196%)
F Converted at 0.2% (0% to 0.7%) with 2 out of 1138 visits 0.52 -36% (-140% to 92%)
G Converted at 0.4% (0.1% to 1.0%) with 4 out of 1123 visits 0.78 +29% (-112% to 145%)
Data is shown at a 95% confidence interval and adjusted for multiple variations using the Bonferroni correction.
At the same time we wanted the customer to make an informed choice. For instance, we decided to soften
the name to “A month or so” and label it with “No term” – making it overall more appealing. We also
decided to use the higher Monthly price with larger font size as a subtle anchor to suggest other plans had
more value and add explicit messages to that effect. Overall we wanted to create a “here are the facts, you
make your own choice” impression.
Test Setup
A Converted at 16% (15% to 16%) with 970 out of 6248 visits p-Value Improvement
H Converted at 17% (16% to 18%) with 1119 out of 6438 visits 0.005 +12% (4% to 20%)
Data is shown at a 95% confidence interval and adjusted for multiple variations using the Bonferroni correction.
A Converted at 6.1% (5.5% to 6.7%) with 380 out of 6248 visits p-Value Improvement
H Converted at 5.5% (16% to 18%) with 355 out of 6438 visits 0.17 -9.3% (-23% to 4%)
Data is shown at a 95% confidence interval and adjusted for multiple variations using the Bonferroni correction.
A Converted at 1.3% (1.1% to 1.6%) with 83 out of 6248 visits p-Value Improvement
H Converted at 1.5% (1.2% to 1.8%) with 97 out of 6438 visits 0.4 +13% (-18% to 44%)
Data is shown at a 95% confidence interval and adjusted for multiple variations using the Bonferroni correction.
Although Luke Wroblewski’s well-known experiments have shown that single page forms perform similarly
to multi-page forms, that may not hold for every kind of multi-step process. In some cases, it may be
better to split a complex process into several steps. In this case, consolidating the choices, sales copy, and
payment on one page was likely more effective.
We believe what worked was a balanced presentation of the choices and their advantages and disadvan-
tages, so the user could make the choice that is right for them. For example, in variation H the higher price
for the 1 month plan was paired with the label “No term” – disadvantage and advantage. Instead of pushing
one choice, we cross-promoted the choices. For example, the 1 month plan says “Don’t want to rush? Save
at least $30 with the 3-month plan”. This states a potential disadvantage (rushing) and suggests another
plan as a better option. We also renamed this plan from “Month” to “A Month Or So”, which may have
provided a time context that more people could related to.
Another important change we made was to interpret the prices for our visitors. Instead of just stating a
price and a saving %, we tried to explain what it would mean for them. For example, the 3-month and
6-month plans say “pays for itself in X months”, putting that price clearer context. Likewise, the 6-month
plan not only states a price but suggests who this plan might best suit: “take your time meeting people”.
Overall, we feel this variation better engaged the visitors’ thought process and left less doubt as a result.
I Fewer Form Fields - The credit card field is one which could be detected automatically.
J Curiosity - Reminding about the profiles which were looked at by the user.
K Social Proof - Hinting at who else signed up today from the area the user searched for.
If you’d like to see a particular test, have an idea or comments about this issue, please send your emails to jakub@linowski.ca.
Thank you again for supporting the project! We hope you enjoyed the issue and learned from our insights. :)
Also a big thanks to Visual Website Optimizer, www.VWO.com, for supporting us by providing Datastories with an awesome a/b testing tool.
LINOWSKI
INTERACTION DESIGN