Tim Wilson's presentation at eMetrics San Francisco in April 2013 covering an approach for analysts to establish and manage a well-defined process for hypothesis discovery, prioritization, and testing to help pivot companies to be data-driven -- to orient towards a culture of validated learning.
13. Align on goals and KPIs
Discover hypotheses across the company
Assess the hypotheses and your approach
Prioritize the hypotheses
Test the highest priority hypotheses
Act on the results
Learn for the future
/emetrics @tgwilson
14. Align on goals and KPIs
Discover hypotheses across the company
Assess the hypotheses and your approach
Prioritize the hypotheses
Test the highest priority hypotheses
Act on the results
Learn for the future
/emetrics @tgwilson
18. Magic questions.
1. What are we trying to achieve?
2. How do we know if we’ve done that?
/emetrics @tgwilson
19. A Process: Goals and KPIs
Two Magic − What are we trying to achieve?
Questions − How will we know if we’ve done that?
Establish − Open-ended questions
Appropriate − “What?” and “How?” (not “Why?”)
Measures − The power of restatement
− Historical data
Establish
− Back-of-the-napkin creativity
Targets
− Bracketing
/emetrics @tgwilson
20. Online Revenue Engagement Reach
Drive online revenue through the presence of xxxx Increase overall site engagement Increase the reach of [site].com
Conversion Rate Lift +14.3% New Profiles Created 2,707 SEO Visits to xxxx 116
Visits that interacted with xxxx 7.08% Total Profiles created since launch 3,391
20
250
Visits that did not interact with xxxx 6.19% 200 15
15.0% 150
10
10.0% 100
5.0% 50 5
0.0% 0
3/10 3/26 3/10 3/26 0
3/10 3/26
Overall conversion Projected rate Sitewide
rate for the site without xxxx % Lift % Interacting w/ xxxx 13.8%
+3.19% +3.14% +1.6% Visits that interacted 36,605 Social Shares of xxxx 251
Visits that did not interact 229,125 50
Avg. Order Value Lift +2.7% 20.0% 40
Visits that interacted with xxxx $23.37 15.0%
30
20
Visits that did not interact with xxxx $22.76 10.0%
10
$40 5.0%
0
0.0% 3/10 3/26
$20
3/10 3/26
$0
3/10 3/26
Time on Site Lift +77.6% Social Referral Visits 46
Overall AOV Projected AOV Sitewide
for the site without xxxx % Lift Visits that interacted with xxxx 24:49 8
$23.04 $22.96 +0.3% Visits that did not interact with xxxx 13:59
6
30:00 4
Total Revenue Lift $8,976 25:00
20:00 2
Conversion rate lift only $7,389 15:00
0
AOV lift only $1,389 10:00 3/10 3/26
05:00
AOV lift on incremental orders $199 00:00
3/10 3/26 Refer to the glossary for a detailed definition for each metric.
Annualized Lift (ignoring seasonality) $192,727
/emetrics @tgwilson
21. Align on goals and KPIs
Discover hypotheses across the company
Assess the hypotheses and your approach
Prioritize the hypotheses
Test the highest priority hypotheses
Act on the results
Learn for the future
/emetrics @tgwilson
26. The trick is getting
them articulated
clearly.
/emetrics @tgwilson
27. What we say to marketers
“What is your
hypothesis?”
What they hear
“I’m dragging you
back to the 8th grade
science class you
hated.”
/emetrics @tgwilson
29. Each idea can be
added to your
hypothesis library.
Google Spreadsheet
Excel
uservoice.com
Sharepoint
Lotus Notes
Jira
/emetrics @tgwilson
30. Align on goals and KPIs
Discover hypotheses across the company
Assess the hypotheses and your approach
Prioritize the hypotheses
Test the highest priority hypotheses
Act on the results
Learn for the future
/emetrics @tgwilson
31. To assess the effort
required, you have to
determine the
approach.
/emetrics @tgwilson
34. I believe that…
[some idea]
If I am right, then…
[some action]
To test this, I will…
[tool/technology/approach]
/emetrics @tgwilson
35. What’s in your
company’s toolbelt?
Web Analytics Email Surveys
Voice of the Customer Social Media Analytics
A/B (and MV) Testing Focus Groups
Usability Testing Secondary Research
/emetrics @tgwilson
36. Align on goals and KPIs
Discover hypotheses across the company
Assess the hypotheses and your approach
Prioritize the hypotheses
Test the highest priority hypotheses
Act on the results
Learn for the future
/emetrics @tgwilson
37. It’s not a simple formula.
/emetrics @tgwilson
38. But a formula can
guide the
discussion.
/emetrics @tgwilson
39. Likelihood belief is right H M L
Likelihood change would occur H M L
Likely scale of impact H M L
Alignment w/ business goals H M L
Effort to test (approach) H M L
/emetrics @tgwilson
40. Align on goals and KPIs
Discover hypotheses across the company
Assess the hypotheses and your approach
Prioritize the hypotheses
Test the highest priority hypotheses
Act on the results
Learn for the future
/emetrics @tgwilson
42. Align on goals and KPIs
Discover hypotheses across the company
Assess the hypotheses and your approach
Prioritize the hypotheses
Test the highest priority hypotheses
Act on the results
Learn for the future
/emetrics @tgwilson
52. We tested…
[the idea]
As a result, we…
[took the action]
Resulting in…
[quantified result]
/emetrics @tgwilson
53. Align on goals and KPIs
Discover hypotheses across the company
Assess the hypotheses and your approach
Prioritize the hypotheses
Test the highest priority hypotheses
Act on the results
Learn for the future
/emetrics @tgwilson
55. Hypothesis tested Action taken Deeper learning
Hypothesis tested No action taken
Hypothesis tested Action taken
Hypothesis tested Action taken Deeper learning
Hypothesis tested No action taken Deeper learning
Hypothesis tested Action taken
Hypothesis tested Action taken
Hypothesis tested Action taken
/emetrics @tgwilson
57. Align on goals and KPIs
Discover hypotheses across the company
Assess the hypotheses and your approach
Prioritize the hypotheses
Test the highest priority hypotheses
Act on the results
Learn for the future
/emetrics @tgwilson
58. clearhead.me/emetrics
This presentation
Guide: getting to great KPIs
Hypothesis library template
Approach options self-audit template
Recommended reading
/emetrics @tgwilson