Bringing you Tips on Running....

A|B Testing | Optimization|Experimentation

  • Experimentation Expert

WARNING | Covid-19 Causes Mass Panic! Don't Let Your Test = ERROR !


(Better view on Desktop)


1 UH OH ERROR!

2 We all are seeing the mass hysteria crowding around the streets and our grocery stores. So

3 I can image the panic from your leaders wondering what's happening with your tests and

4 can they still rely on the data and observed trends.

5 Look at the quick "covid_19 code" on how to address your leaders on reporting pressure.

6 #DONTPANIC

7

8 Follow in this order :-).."Warning to my software engineers, this code is NOT REAL :-)"

9

10

11 ENTER_

12 RUN <covid_19>

13 IF "Test A" confidence < 95%

14 THEN DO NOT Report_PostAnalysis WHY;

15 Answer= "It is poor practice to peek at data and report on unrealistic numbers. Samples

16 are already statistical observations of your population. A population you once had a

17 strong confidence in on behavior is now behaving like your KPI was bit by data rabies. "

18

19 "IF you are unsure on how to understand the deviation of your mean for your typical

20 variants when running tests for your business THEN RUN an A|A test to observe how the

21 data flows when isolated into two equal groups within a sample of your population."

22

23 "IF "Test A" start_date < MMDDYYYY Enter observed D-day for your data. Or the point

24 that the business notices an impact to KPIs THEN freeze and isolate the data to have

25 comparable pre and post apocalypse analyses"

26

27 ATTENTION LEADERS ASK YOUR EXPERIMENTATION TEAM <What Happened>

28 RUN <What Happened>

29 "Under no circumstances should you make decisions without proper analysis."

30 IF "Test A" = inconclusive

31 <check sample size was met> IF sample_met=no THEN RUN = continue

32 <check analytics> IF "TAG is misfiring" THEN RUN = discontinue

33 <check variants> IF "probability of test visibility is low" THEN modify = variants

34

35 RUN <solutions> RETURN "options"

36 <COVID-19 related tests>

37 <empathy driven tests>

38 < boost products that are on the list of necessities to increase their conversion>

39 "I hope you enjoyed, see you next time :-) "

40

41 END_



If you have any fun topics and ideas you would love for me to cover, not necessarily in this style, drop a comment!

9 views0 comments