Upload
equilibria-inc
View
222
Download
0
Embed Size (px)
Citation preview
BEYOND THE CLICKSlideShare as a Lead Generation Tool – a Case Study
© 2016 Equilibria, Inc. All Rights Reserved
FREE process
map inside!
Contents
Introduction
The Challenge
Data Mining – Six Sigma
Insights
Slide 03
Slide 12
Slide 19
Slide 58
Will this Journey Lead Towards the Light of More Sales…
Introductions
Image Credit: Elite Motor Performance (UK)
…or Into a Bottomless, Time-Wasting Pit?
About5 | INTRODUCTIONS
www.eqbsystems.com
The Journey Began With a Hunch6 | INTRODUCTIONS
2009
● My Intro to SlideShare● Began Uploading
Content
● LinkedIn Acquired SlideShare
2012 2013
● Received Congratulatory Email from SlideShare (content in top 1% viewed)
www.eqbsystems.com
So I launched an investigation…7 | INTRODUCTIONS
www.eqbsystems.com
and was shocked by what I discovered!8 | INTRODUCTIONS
www.eqbsystems.com
A New Metric…Outbound Click9 | INTRODUCTIONS
www.eqbsystems.com
10 | INTRODUCTIONSIm
age
Cred
it: m
irror
.co.
uk
www.eqbsystems.com
Can an increase in
lead to an increase in online sales?
11 | INTRODUCTIONS
THE MILLION DOLLAR QUESTION
OUTBOUND CLICKS
www.eqbsystems.com
Increase Outbound Clicks from SlideShare to Website
The Challenge
Online Sales Process13 | THE CHALLENGE
Find Content Content Engagement
VIEWER CONSUMER
www.eqbsystems.com
CUSTOMER
PURCHASE
Content
Engagement Options
Outbound Links
Call-to-Action
Inbound Links
1
2
3
4
5
14 | THE CHALLENGE
Scre
ensh
ot c
irca
1st Q
tr 2
014
Anatomy of a SlideShare Post - Engagement
2
1 3
45
www.eqbsystems.com
Description
Viewer Comments
Total Page #s
Number of Views
Title
15 | THE CHALLENGE
Anatomy of a SlideShare Post - StatsSc
reen
shot
cir
ca 1
st Q
tr 2
014 6
7 8
9
10
6
7
8
9
10
www.eqbsystems.com
16 | THE CHALLENGE
Anatomy of a SlideShare Post - DataSc
reen
shot
cir
ca 2
nd Q
tr 2
014
www.eqbsystems.com
17 | THE CHALLENGE
Overcoming “The Challenge”
Consumer activity on SlideShare can help us figure out which products & services to
create.
Not a lot of data to work with
Increasing outbound click can be difficult
Investigate, but don’t spend a lot of time
Not convinced SlideShare is a viable lead generator
MY ARGUMENT
www.eqbsystems.com
18 | THE CHALLENGE
Moving Forward - a New Project
2010
● My Intro to SlideShare● Began Uploading
Content
● LinkedIn Acquired SlideShare
2012 2013
● Received Congratulatory Email from SlideShare (content in top 1% viewed)
2014
www.eqbsystems.com
● Start project to increase outbound clicks
Using the DMAIC Methodology for Insights
DATA MINING – SIX SIGMA
6σ
20 | DATA MINING - SIX SIGMA
DEFINE1
MEASURE2
ANALYZE3
IMPROVE4
CONTROL5
21 | DATA MINING - SIX SIGMA
• Content Developer
• SlideShare• Scan Search
Results
• Email• Comment• Download• Social Share
• Consumer to Lead conversion
?• Select
Search Result w/ Direct or Indirect Link
• Keyword Search
• Confirm content addresses need
• Want more info
PROCESS
View Content on
SlideShare
Search for Content Online
Respond to Content
Click to EQBsystems.c
om
Defining the Path to Y – SIPOC (current state)SUPPLIERS INPUTS OUTPUTS CONSUMER
S
www.eqbsystems.com
DEFINE | MEASURE | ANALYZE | IMPROVE | CONTROL
22 | DATA MINING - SIX SIGMA
WHO Is Our Consumer?
www.eqbsystems.com
DEFINE | MEASURE | ANALYZE | IMPROVE | CONTROL
23 | DATA MINING - SIX SIGMA
WHERE Is Our Consumer?
www.eqbsystems.com
DEFINE | MEASURE | ANALYZE | IMPROVE | CONTROL
24 | DATA MINING - SIX SIGMA
WHAT MAKES Our Consumer Respond?
www.eqbsystems.com
DEFINE | MEASURE | ANALYZE | IMPROVE | CONTROL
25 | DATA MINING - SIX SIGMA
HOW/WHYDoes Our Consumer Respond?
www.eqbsystems.com
DEFINE | MEASURE | ANALYZE | IMPROVE | CONTROL
26 | DATA MINING - SIX SIGMA
Voice of Customer – Extrapolation
FUNCTIONALRELEVANT SUPPORTIVE
• Use relevant tags/keywords
• Content should match my search query
• Provide content that meets my needs
• Make it easy for me to share content
• Include content to explain images
• Include links only when they are active
• Keep your promises/guarantees
• Respond quickly to my inquiries/comments
• Give me multiple ways to contact you
www.eqbsystems.com
DEFINE | MEASURE | ANALYZE | IMPROVE | CONTROL
Incorrect/missing tagsIncorrect/missing keywords
27 | DATA MINING - SIX SIGMA
Error / Defect IdentificationDefective Content has:
www.eqbsystems.com
DEFINE | MEASURE | ANALYZE | IMPROVE | CONTROL
Missing or sub-par descriptionLess than four pages in length
Punctuation & grammatical errorsMissing or sub-par call-to-action
Low views (not evergreen)Non-searchable title
Inactive linksIncorrect information
28 | DATA MINING - SIX SIGMA
Defining Y – Outbound Click Response
-1 0 1 2
25
20
15
10
5
0
OUTBOUND CLICKS (Y)
FREQ
UENC
Y
HISTOGRAM OF OUTBOUND CLICKS (FEB.2014)
Mean 0.1724StDev 0.5391N 29
NORMAL 90%content with zero outbound clicks!
Examine content with outbound
clicks. What do they
have that others don’t?
www.eqbsystems.com
DEFINE | MEASURE | ANALYZE | IMPROVE | CONTROL
29 | DATA MINING - SIX SIGMA
Defining the Project Scope
Increase outbound clicks by 15% from our accounting-related content on SlideShare
to our online store in 4 months.
GOAL
Specific Measurable
Attainable
Relevant Time Based
www.eqbsystems.com
DEFINE | MEASURE | ANALYZE | IMPROVE | CONTROL
30 | DATA MINING - SIX SIGMA
iNSANiTY Check
I suspect that with an increase in the performance of each x it will lead to an increase in Y, giving us more data to test.
Any increase in Y needs to be viewed as a success.
SPECIFICYes, narrowed down to accounting-related account info
MEASURABLEQuantitative, but discrete data - can’t guarantee accuracy and precision of data
ATTAINABLEA very difficult task with an industry avg. of 3% outbound click
RELEVANTYes, but may not see great improvements
TIME-BASEDYes, since the scale of this is small
www.eqbsystems.com
DEFINE | MEASURE | ANALYZE | IMPROVE | CONTROL
6σ
31 | DATA MINING - SIX SIGMA
DEFINE1
MEASURE2
ANALYZE3
IMPROVE4
CONTROL5
Y = outbound click (count)x = factors that contribute to outbound clicks
32 | DATA MINING - SIX SIGMA
Measuring Relationship Between Y and x
Y= f(x)
www.eqbsystems.com
DEFINE | MEASURE | ANALYZE | IMPROVE | CONTROL
Length
Download
Embed Call-to-Action
Timing
Title
Description
Tags
Keywords
The past year has only yielded 5
outbound clicks to the online store from acct-
related content on SlideShare.
33 | DATA MINING - SIX SIGMA
Determining the “x” Factors
www.eqbsystems.com
DEFINE | MEASURE | ANALYZE | IMPROVE | CONTROL
DiscreteCount
DigitalMeasurement
34 | DATA MINING - SIX SIGMA
Download Data - Description
www.eqbsystems.com
DEFINE | MEASURE | ANALYZE | IMPROVE | CONTROL
1. Examine entire population of content to determine number of downloads.
2. Calculate a weekly ratio of downloads to total views per content.
3. Express as a % up to three decimal places.
35 | DATA MINING - SIX SIGMA
Download Data - Collection
SAMPLE
POPULATION
sampling
www.eqbsystems.com
DEFINE | MEASURE | ANALYZE | IMPROVE | CONTROL
Process appears to be in control for predicting the download response ratio.
36 | DATA MINING - SIX SIGMA
Download Data - Representation
www.eqbsystems.com
DEFINE | MEASURE | ANALYZE | IMPROVE | CONTROL
37 | DATA MINING - SIX SIGMA
iNSANiTY Check
Now that I think we can predict our download data (process is stable)...
In 6σ projects, goal is to have continuous data to determine if a statistical correlation between Y
and x exists.
● Current process allows defective content to pass through● Creates skewed data● Outbound click data only available as of May 2013
www.eqbsystems.com
DEFINE | MEASURE | ANALYZE | IMPROVE | CONTROL
Is there a statistical correlation between outbound click, Y and Download Data (xD)?
6σ
28 | DATA MINING - SIX SIGMA
DEFINE1
MEASURE2
ANALYZE3
IMPROVE4
CONTROL5
Y = outbound click (count)xD = download response (ratio)
Is Y = f(xD) + f(x2) + f(x3) + f(x4) + f(x5) ?
Prove/disprove the effect of each x on Y through statistical testing.
39 | DATA MINING - SIX SIGMA
Analyzing Relationship Between Y and x
www.eqbsystems.com
DEFINE | MEASURE | ANALYZE | IMPROVE | CONTROL
STEPS
1. Define Ho and Ha
2. Set level of significance: α = 0.05
3. Select test based on type of
data.
4. Calculate p-value from selected
test.
Image: imgarcade.com
40 | DATA MINING - SIX SIGMA
Test Hypothesis
www.eqbsystems.com
DEFINE | MEASURE | ANALYZE | IMPROVE | CONTROL
41 | DATA MINING - SIX SIGMA
Hypothesis Theory or Fact?
p-value Ho Ha
< α
> α
CRITICAL X
www.eqbsystems.com
DEFINE | MEASURE | ANALYZE | IMPROVE | CONTROL
Ha Seeing number of downloads has an effect on Viewer outbound clicks.
Seeing number of downloads has no effect on Viewer outbound clicks.
Ho
Accept Ho ?
42 | DATA MINING - SIX SIGMA
Test Results – Logistic Regression
www.eqbsystems.com
DEFINE | MEASURE | ANALYZE | IMPROVE | CONTROL
Reject Ha ?
NOT READY to affirmatively state that we have no critical x’s.
Continue collecting data until all 29 pieces of content are optimized and tested against our new internal QA Checklist.
Y = f(xD) + f(x2) + f(x3) + f(x4) + f(x5)
? ? ? ? ?
43 | DATA MINING - SIX SIGMA
iNSANiTY Check
THEREFORE…
www.eqbsystems.com
DEFINE | MEASURE | ANALYZE | IMPROVE | CONTROL
6σ
44 | DATA MINING - SIX SIGMA
DEFINE1
MEASURE2
ANALYZE3
IMPROVE4
CONTROL5
New Process
45 | DATA MINING - SIX SIGMA
+
New People
+
New Tools
www.eqbsystems.com
DEFINE | MEASURE | ANALYZE | IMPROVE | CONTROL
High Level Process Map: Lead Generation via SlideShare ContentSE
O An
alyst
Cont
ent D
evelo
per
Viewe
r
Search for Content
Solutions Online
Develop New/Optimize
Existing Content
Provide Feedback on
Content Description
Upload/Re-upload Content
to SlideShare
Analyze Search Terms Used in
Google Analytics
Self-Identify Business Pain
Point(s)
Analyze Content Interaction Data
in SlideShare PRO
Download Content
Equilibria, Inc.Alicia Butler Pierre
NOTE: there is not a separate swimlane for SlideShare PRO due to real estate constraints
View Content on SlideShare
Add Content Description with
Keywords
Send Draft of Content
Description to SEO Analyst
START
Respond to Content?
Select Result with Link to
Content?
Response(s) Type?
Share Content
Email Content
Comment
Mark as Favorite
Click-Through to Online Store
Share Data Trends with
Content Developer
Lead to Client Conversion
Process
no
yes
Yes
2
3
START
END
no(Lost Viewer Opportunity
possibly due to low search ranking results)
yes
3
3
3
3
3
1
1
2
www.eqbsystems.com
(Current State)
www.eqbsystems.com
Not sure which x’s are critical, so:
1. Optimize accounting-related content first using Error/Defect checklist
2. Monitor effect on download ratio, xD
3. Test effect on outbound click, Y
4. Confirm the critical x’s
5. Optimize remaining content
48 | DATA MINING - SIX SIGMA
Content Optimization Plan
www.eqbsystems.com
DEFINE | MEASURE | ANALYZE | IMPROVE | CONTROL
49 | DATA MINING - SIX SIGMA
Content Optimization Schedule
www.eqbsystems.com
DEFINE | MEASURE | ANALYZE | IMPROVE | CONTROL
50 | DATA MINING - SIX SIGMA
Content Optimization Example
BEFORE
www.eqbsystems.com
DEFINE | MEASURE | ANALYZE | IMPROVE | CONTROL
x2: Call-to-Action x1: Content Length > 4pgs
x3: Corresponding Blog/Video Post
for Embed Viewing
xD: Content can be downloaded
x5: Content can be emailed
51 | DATA MINING - SIX SIGMA
Content Optimization Example
AFTER
www.eqbsystems.com
DEFINE | MEASURE | ANALYZE | IMPROVE | CONTROL
52 | DATA MINING - SIX SIGMA
What Happens
?How Bad?
What are the
Causes?How
Often?Are We
In Control?
Murphy’s Law – Preparing for Failure
What Can Go Wrong?
What Action Do We Take?
www.eqbsystems.com
DEFINE | MEASURE | ANALYZE | IMPROVE | CONTROL
53 | DATA MINING - SIX SIGMA
iNSANiTY Check
We are heading in a positive direction
We should see an upward shift in the mean of each control chart
● Outbound click rate increased from 5 to 23● Hired four additional contractors to ensure a seamless process● Developed a control plan and FMEA to manage risk
www.eqbsystems.com
DEFINE | MEASURE | ANALYZE | IMPROVE | CONTROL
6σ
54 | DATA MINING - SIX SIGMA
DEFINE1
MEASURE2
ANALYZE3
IMPROVE4
CONTROL5
BEFORE Mean shifted up by 132.6%. Ignore error
message.
55 | DATA MINING - SIX SIGMA
Process In Control, but Still No Critical xAFTER
www.eqbsystems.com
DEFINE | MEASURE | ANALYZE | IMPROVE | CONTROL
APP ARP CTR VAS0
3
6
9
12
15
2
0
21
4
0
12
14
Before After
Y: outbound click from SS content to online store By this point, only 2 out
of the 4 pieces of accounting-related content were optimized. It makes sense that this is also where we see the greatest improvement in outbound click.
56 | DATA MINING - SIX SIGMA
Total Before
Total After % Increase
5 30 600%
Y: Before & After Content Optimization
www.eqbsystems.com
DEFINE | MEASURE | ANALYZE | IMPROVE | CONTROL
The following Control Planning tools ensure the process runs as smoothly as possible:
57 | DATA MINING - SIX SIGMA
Today’s Improvement Becomes Tomorrow’s Standard
Process Maps Work Instructions Control Plan Job Aids FMEA
www.eqbsystems.com
DEFINE | MEASURE | ANALYZE | IMPROVE | CONTROL
Flow in FluxTMINSIGHTS
Accomplishments:• Big Y vs. Little Y
(narrow)• Big Y = ↑ online sales• Little Y = ↑ outbound
click• Defined project scope• Project’s Y
identification• Key stakeholder
identification• Internal and external
customer identification
Deliverables:• Project Charter• Gantt Chart• SIPOC• High Level Process
Map• VOC Matrix• COPQ Matrix
Discoveries:• Countries with
highest viewership• Broken links in online
store• SlideShare changed
its embedded link policy
Accomplishments:• Defined process
baseline performance• Identified CS process
gaps & bottlenecks• Identified 5 potential
x’s to answer: is y=f(x) ?
• Identified Null & Alternative Hypotheses
• Detailed Operational Definitions for data collection on each x
Deliverables:• Value Stream Map (CS)• Affinity Diagram• Fishbone Diagram• Data Collection Plan• MSA• Control Charts• DPMO & Sigma Level• Preliminary AnalysisDiscoveries:• Some data easier to
retrieve than others
Accomplishments:• Ran statistical tests on
each x; α set to 0.05
Deliverables:• Value Stream Map (IS)• Hypothesis Testing• Data Collection Plan
(completed)
Discoveries:• Continuous data works
best for statistical testing• All of our x and Y data is
discrete and considered an integer/ratio for statistical test selection
• More Y data needed in order to “trust” initial results of critical x determination
• SlideShare PRO’s analytical tool should be tested against Google Analytics through use of tracking content URLs
Accomplishments:• Conducted pilot test on 2
out of 4 pieces of accounting-related content (VAS & CTR)
• Created a J-I-T content optimization schedule
Deliverables:• Value Stream Map (FS)• FMEA• Content Optimization
ScheduleDiscoveries:• Outbound click rate
increased on the 2 pieces of optimized content
• New FS process has more steps and more people, yet yields a shorter cycle time
• FMEA exposed process vulnerabilities of working with independent contractors
• Disaster Recovery planning tantamount to process success
Accomplishments:• Identified corrective and
preventive actions for technological breakdowns and labor shortages
• Increased frequency for quality assurance checks
• Traced first client’s journey from SlideShare content!
Deliverables:• Control Plan• Visual Process Metrics• Change Management
Plan• Work Instructions & Job
Aids• Project Transition PlanDiscoveries:• The SlideShare app does
not have same functionality as the website – outbound links do not work regardless of location
• Acquisition activity makes SlideShare, as a company, vulnerable to platform changes
Summary of Activities & DeliverablesDEFINE MEASURE ANALYZE IMPROVE CONTROL
www.eqbsystems.com
60 | INSIGHTS
Bookkeepers ● Office Managers ● Call Center Employees ● Virtual/Offshored/Outsourced Employees ● D-I-Y Entrepreneurs
Customer Profile
www.eqbsystems.com
United StatesIndia
Canada
Philippines
United Kingdom
Real-Life Success Story
www.eqbsystems.com
Noticed same
consumer downloading
content.
1
Thanked consumer
and offered to give free
consultation.
2Consumer
accepted offer for free
consultation.
3
Consumer became a Customer!
4
61 | INSIGHTS
62 | THE CHALLENGE
From Special Project to Operations
2010
● My Intro to SlideShare● Began Uploading
Content
● LinkedIn Acquired SlideShare
2012 2013
● Received Congratulatory Email from SlideShare (content in top 1% viewed)
2014
www.eqbsystems.com
● Start project to increase outbound clicks
TODAY
● Ongoing optimization & new content creation
Key Insights:
63 | INSIGHTS
© josephgiitterphotography.com/
Vigilance is Key
www.eqbsystems.com
1. Have sufficient data to apply Six Sigma testing
2. Perform Quality Assurance weekly (minimum)
3. Pay attention to social network acquisition activity 4. Approach social media data as a snapshot in time
The Single, Most Important Insight…When It Comes to Working with Digital, Social Media Data
www.eqbsystems.com
© Copyright Netflix.
Stay on Top of Your Game
www.eqbsystems.com
SlideShare no longer reports outbound click data!
About Me
I AM…
Alicia Butler Pierre@EfficiencyEngr
www.eqbsystems.com
● Founder – Equilibria, Inc.● Engineer & Lean Six Sigma
PractitionerI AM NOT…
● A social media or Minitab guru● A behavioral or social scientist
THANK YOU FOR VIEWING!
Contact Us for more questions.
We condensed a lot of information for this presentation.
www.eqbsystems.com
@EquilibriaInc@Equilibria, Inc @EQBsystems
©Copyright 2016. Equilibria, Inc.