AB Update (2022-11-23)

Note: You’ve found an archived AB Update. Search for more recent AB Updates (…you can start AB Testing & Experimentation Update (2022-12-15) )

AB Update Email || Subject: AB Testing & Experimentation update | Sent 2022-11-23

Hello Team—Before the Thanksgiving holiday kicks off, we wanted to provide a quick update regarding the latest AB Experimentation results for November's tests.

Summary of November’s AB Tests—

(1) Winner:
- Yankee Checkout CC Expiration Text (MMYY)
(2) Tests Trending Positive:
- Yankee Global NAV Carousel Stripe Banner
- Yankee NAV L1 “Shop By” to “New Products”
(3) Learners:
- Graco NAV Shop All redirect to Landing Page (vs Journey Page)
- Yankee Auto-Ship Toggle Location (global change)
- Yankee PDP Fragrance Notes

Results & Details—

Winner ⭐️ Yankee Checkout CC Expiration Text

Hypothesis: By adding allowed helper text near credit card expiration, we will better convey to users allowed format, which will decrease friction and increase conversions.

Results: after 18 days, variant V1 (with MMYY Helper Text) is a Winner!
This minor edit goes to show that even small tweaks within checkout can have huge impacts!
Engagement: V1 saw a 0.75% improvement in users clicking on the Submit Order button with 96% stat sig.
Conversions: V1’s Transaction Rate improved 0.48% (with 94% stat sig) and was at parity with control for Mobile. This could be due to the fact that autocomplete on mobile phones is likely more prevalent than on Desktop browses.
Recommendation: End test and implement the winning V1 variant.

See this test’s wiki page for the latest!
https://newellbrands.atlassian.net/wiki/spaces/ABT/pages/3427500041/Winner+Yankee+Checkout+CC+Expiration+Date+MMYY+helper+text+ABT-166+CXP.023

Trending Positive: Yankee Global Carousel Stripe

Hypothesis: By making the Global Carousel Stripe static with one consistent message or by removing it entirely, we will reduce friction, thus increasing carousel engagement and purchases.

Results:
- After 7 days of run time, both V1 (Static) & V2 (banner removed) are Winning.
- In just 7 days, this test generated $25k in revenue above the control.
Engagement: Control saw more clicks to the carousel with stat sig. V1 saw 22% fewer clicks and V2 saw virtually no clicks due to the carousel element being removed. PDP page views are seeing stat sig improvements under the two variations with a 1.16% lift for V1 and 0.75% increase for V2.
Conversions: Transaction rate is showing great improvements for V1 and V2. V1 shows a 1.52% lift in transaction rate at 95% stat sig and V2 is showing a 1.88% lift with 97% stat sig.
Recommendation: Early results are very promising, but it is still too early to tell. Sustain test for another week for more definitive results.

Trending Positive: Yankee NAV L1 “Shop By” to “New Products”

Hypothesis: By redirecting users from the “Shop By” section to the “New” landing page we will be resonating more with users, increasing engagement and conversions.

Results: after 13 days, V2 (“New Products”) is Winning for both L1 Click Engagement & Transaction Rate metrics.
Engagement: V2 generated the largest lift in L1 nav clicks with 92% more clicks on L1 at 100% stat sig. V1 ("New”) & V3 ("Shop New") also saw increases in L1 Clicks.
V1 saw a 39% increase and V2 41% increase (both stat sig). L2 NAV Clicks went down for all variations but with the increase in Clicks to L1 element, users are finding what they need through the variation without having to drill down to L2.
Conversions: V2 saw a 10% lift in transaction rates with 89% confidence, nearly stat sig. V1 and V3 saw stat sig improvements in AOV, whereas V2 showed improvement but without stat sig. However, when looking at potential revenue lift across the three variations, V2 shows the most potential to increase revenue. Within the 13 days that this test has run, V2 generated $38k in revenue uplift over the control.
Recommendation: implement V2 after running for another 1-2 days to verify test results.

Learner: Graco Global Shop All Nav

Hypothesis: By redirecting Car-Seat and Stroller navigation clicks to Shop-All-pages instead of Journey pages, we will allow users to start browsing and shopping earlier in the process, thus increasing PDP pageviews, Add to Carts, and Purchases.

Results: after running for 17 days, this test’s results are inconclusive.
Engagement: V1 (Shop All) Links pageviews increased 0.35% with 94% stat sig. Clicks on the Stroller-L1 & Car-Seats-L1 links showed some improvements but lacked stat sig.(which isn’t entirely surprising as there are no visible differences between V1 and the control that would change user behavior).
Conversions: Transactions Rates & AOV saw some uptick but lack stat sig.
Recommendation: End test and consider implementation based on preference.

Learner: Yankee Global Auto-Ship Toggle Location

Note: Final test results may change as this test was recently restarted—expect final results by the end of November.
Hypothesis: By moving the Auto-Ship toggle next to the sort drop down, we will increase toggle interaction and increase purchases with subscriptions.

Results: after 13 days of run time, the test results are inconclusive.
Engagement: Auto-ship toggle clicks in V1 (Toggle Top) were down 21% with stat sig. However, interestingly auto-ship radio clicks on PDP were up 15% with 91% stat sig. Potential explanation for this is that since the toggle is not in the sidebar filter, users who sought auto-ship and clicked the toggle had a higher intent to transact with auto-ship whereas users may hit the toggle in the sidebar filter regardless of intent.
Conversions: Transaction rate is down for V1 with control winning nearly at stat sig, but there is indication that AOV could increase under V1 with a 0.1% increase in Average Order Quantity (however, not at stat sig).
Recommendation: Conclude test and move on to more fruitful AB tests.

Learner: Yankee PDP Fragrance Notes

Hypothesis: By making the fragrance notes more visible, we can provide more information to the users, increase add to bag clicks and conversions.

Results: this ran for 14 days, and results are inconclusive.
Engagement: V1 (Expanded) saw a 1.26% increase in Add to Bag button clicks on PDP with 99% stat sig. However, progression through the checkout funnel saw an overall decrease. Cart page visits saw a 0.69% decrease with stat sig and subsequent funnel step page views saw a decrease.
Conversions: V1 transaction rates saw a 1.32% decline with stat sig. AOV saw a slight uptick but is essentially at parity with control. It is unknown why ATB clicks went up, but users did not follow through with their purchases. Filtering down to different user demographics such as device or new/returning visitors did not yield additional insights.
Recommendation: Conclude test and iterate on new test ideas around fragrance notes.

Coming Soon… December’s upcoming AB Tests

We’ve worked with Yankee, Graco, and our Tech teams to arrive at some low-risk tests as we close out this year’s testing in December.

Special thanks to our AB Partners…

We’re already seeing some impressive numbers including a $429.17K uplift resulting from Yankee & Graco’s experiments which we kicked off in July of this year. As you can see your support goes a long way—keep those ideas coming!

Have a test idea? Submit your experiment hypothesis today @ Bit.ly/abRequest or Slack us!
We’re happy to discuss test ideas and address any questions you may have!

Have a great Thanksgiving everyone! 🦃
-Alex