Embed Size (px)
DESCRIPTION
Citation preview
Social Media Analytics - What, Why and How A Presentation to MESH Marketing October 22, 2009Katie Delahaye [email protected]:/kdpaine.blogs.comMember, IPR Measurement Commissionwww.instituteforpr.org
Eyeball counti
ngHITS Engag
ement
MSM Online Social Media
A measurement timeline
Old School Metrics
AVEsEyeballsHITS (How Idiots Track Success)Couch Potatoes# of Twitter Followers (unless you’re a celebrity)# of Facebook Friends/Fans (unless they donate money)
Page 3
Signs that it’s the end of measurement as we know it 1. 48% of respondents to a PRWeek study said they
were moving $$ out of advertising budgets into Social Media. Only 18% said they were taking $$ away from PR.
2. Procter & Gamble is now paying for engagement, not eyeballs
3. Sodexo cut $300K out of its recruitment budget using Twitter
4. Immunize BC measured SM success via share of discussion, increased awareness and shots given
5. BMC Software measures communications effectiveness based on contribution to EPS
6. HSUS generated $650,000 in new donations from an on-line photo contest on Flickr
7. The Red Cross measures the effectiveness of Twitter via lives saved and property lost
8. IBM predicts the ends of advertising as we know. Also receives more leads, sales and exposure from a $500 podcast than it does from an ad
9. 11 Mom’s turned around Wal-Mart's image and delivered measureable increases in sales.
Page 5
Old School 21st Century
You are a party planner, not a communicator
The definition of timely has changedThe definition of reach has changed
GRPs & Impressions are impossible to count (an irrevelvant) in social media
The definition of success has changedThe answer isn’t how many you’ve reached, but how those you’ve reached have responded Page 6
Old School PR 21st Century Role of PR
Social Media renders everything you know about measurement obsolete
Goals for Social Media
1.Marketing/leads/sales/2.Mission/safety/civic
engagement3.Relationship/reputation/
positioning To fix this Or get to this
Goals drive metrics, metrics drive results
8
Reputation/Relationships
Relationship scores
Recommendations
Positioning
Engagement
Get the word out
% hearing
% believing
% acting
Sales
Engagement Index
Cost per customer
acquisition
Web analytics
Sales leads
Marketing Mix Modeling
Goal
Metrics
Change the conversation, improve your reputation
Improve your reputation
Listen first, then respondStop doing stupid things
Negative coverage over time
Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr2006 2007 2008
0
5
10
15
20
25
21
2
5
21 1
4 42
1 12 2 2
1
10
9
4
1418
21
12
10
15
14
7
26
2
10
4
12
2
3 1
1
1
4
2
2
1
2
5
3
2
2
2
Entr
ies
Correlation exists between traffic to the ASPCA web site and the organization’s
overall media exposure
-
100,000
200,000
300,000
400,000
500,000
600,000
700,000
0
50,000,000
100,000,000
150,000,000
200,000,000
250,000,000
300,000,000
350,000,000
Web
Sit
e Vi
sito
rs
Expo
sure
Overall Exposure
Web Traffic
Tying activity to development/marketing goals
0
50,000,000
100,000,000
150,000,000
200,000,000
250,000,000
300,000,000
350,000,000
Exposure
$0
$200,000
$400,000
$600,000
$800,000
$1,000,000
$1,200,000
$1,400,000
$1,600,000
$1,800,000
Donations
Overall exposure
Online donations
12
Goals, Actions and Metrics Goal Action Output Metric Outtake
MetricOutcome Metric
Increased on-line reservations
Revamp website
Amount of content on web site
% perceiving state as a destination
% increase in web traffic and reservations
#1site for visitors to NH
Increase staffing and resources for communications
Increased exposure of “visit NH” message
Increased perception of NH as an an extreme destination
% increase in agreement with the statement
Website is preferred site for information
Add content, features to web site, keep up to date
% increase in traffic
% agreeing with the statement
# 1 rankings, and time spent on site
The 7 steps to Social Media ROI
1. Define the “R” – Define the expected results?
2. Define the “I” -- What’s the investment?
3. Understand your audiences and what motivates them
4. Define the metrics (what you want to become)
5. Determine what you are benchmarking against
6. Pick a tool and undertake research7. Analyze results and glean insight,
take action, measure again
Step 1: Define the “R” Why Social Media?
What return is expected? – Define in terms of the business or mission.
What problems is Social Media supposed to solve?
What were you hired to do? What difference are you expected to make?
If you are celebrating complete 100% success a year from now, what is different about the organization?
If your Social Media is eliminated, what would be different? 15
Step 2: Define the “I”
What is the investment? PersonnelAgency compensationSenior Staff time Opportunity costRaw costs/hr costs vs material costs.
16
Step 3: Define your audiences and how you impact them
There is no “audience.” There are multiple constituencies Should you blog or Twitter? Don’t ask me, ask your customers Understand your role in getting the audience to do what you want it to do
Raise awarenessIncrease preferenceIncrease engagement
17
Step 4: Define your Key Performance Indicators (KPIs)
18
The Perfect KPIGets you where you want to go (achieves corporate goals)Is actionableContinuously improves your processesIs there when you need it
KPIs should be developed for: Your own propertiesDifferent tacticsOther influential sites
Step 4: Define your Key Performance Indicators (KPIs) carefully because you become what you measure
Cost savingsEfficiency
Cost per message communicatedCost per new lead/customer acquired
Productivity: Increase in employee engagement/moraleLower turnover/recruitment costs
Engagement: Ratio of posts to comments% of repeat visitors% of 5+min visitors% of registrations
Trust:Improvement in relationship /reputation scores with customers and communities (Loyalty/Retention)
Thought leadership: Share of quotesShare of opportunities
Message penetrationPositioning on key issuesImprovement in favorable/unfavorable ratioImprovement in Optimal Content Score (OCS)
19
KPIs for External blogs and other Consumer Generated Media
Share of positioningShare of rants vs. ravesShare of positives/negativesShare of visibilityShare of quotesShare of brand benefits
mentionedTypes of conversationsEngagement – ratio of posts to
comments Optimal content score
Revenue KPIs
Cost savingsCost per click thru, downloads, engagement vs other marketing channelsCost per message communicated vs other channels
Lifetime value of engagementCost per customer acquisition
Engagement metrics
% increase or decrease in unique visits In the past month, what % of all sessions represent more than 5 page views % of sessions that are greater than 5 minutes in duration % of visitors that come back for more than 5 sessions % of sessions that arrive at your site from a Google search, or a direct link from your web site or other site that is related to your brand % of visitors that become a subscriber % of visitors that download something from the site % of visitors that provide an email addressRatio of posts to comments
Courtesy of Eric Peterson
For all institutions, most postings were simply making an observation or distributing media.
Page 23
3
6
1
1
7
36
1
29
5
15
14
2
16
1
2
12
7
2
6
2
24
787
3
2
203
12
12
46
11
1
3
2
1
4
1
4
3
6
2
1
13
2
2
1
13
2
6
18
4
1
1
5
35
3
17
2
8
9
1
1
1
0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%
Acknowledging receipt of information
Advertising Something
Answering a question
Asking a question
Augmenting a previous post
Calling for action
Disclosing personal information
Distributing media
Expressing criticism
Expressing support
Expressing surprise
Giving a heads-up
Giving a shout-out
Making a suggestion
Making an observation
Offering an opinion
Playing a game
Rallying support
Recruiting people
Showing dismay
Share of Conversation Types
Arizona State
Michigan State
Penn State
Purdue University
University of Michigan
44.2%
6.5%
30.9%
49.5%
100.0%
100.0%
100.0%
1.6%
53.9%
100.0%
26.9%
23.1%
10.8%
38.7%
72.7%
10.9%
15.5%
46.1%
66.6%
27.3%
35.1%
39.7%
0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%
Acknowledging receipt of information
Advertising Something
Answering a question
Asking a question
Augmenting a previous post
Calling for action
Disclosing personal information
Distributing media
Expressing criticism
Expressing support
Expressing surprise
Giving a heads-up
Giving a shout-out
Making a suggestion
Making an observation
Offering an opinion
Playing a game
Rallying support
Recruiting people
Showing dismay
Share of Engagement by Conversation Type - Institutional Blogs
Arizona State
Michigan State
Penn State
Purdue University
University of Michigan
cx
Share of conversation vs share of engagement
Page 24
2
2
1
2
1
6
5
3
1
1
1
1
1
1
1
1
1
1
2
2
2
1
2
1
2
1
4
2
1
4
2
1
1
4
1
6
7
6
2
2
2
2
1
3
2
3
1
0 2 4 6 8 10 12 14 16 18 20
Faculty
Students
Research, Physical Sciences
Courses
Research, Earth Sciences
Projects, Non -Research
Financials
Alumni Topics
Research, Life Sciences
Staff
Admissions
Legal News
Other
Research, Agriculture
Policies
Institution, Overall
Campus Life
Research, Social Sciences
Share of Subject
Peer 1
Michigan State
Peer 2
Peer 3
Peer 4
15.3%
68.7%
100.0%
4.4%
33.3%
96.8%
28.6%
34.9%
12.5%
43.3%
28.6%
13.0%
38.3%
100.0%
23.6%
66.7%
6.3%
28.6%
20.8%
2.3%
95.6%
33.2%
5.8%
28.6%
100.0%
86.8%
13.0%
31.0%
22.1%
3.2%
71.4%
43.5%
18.8%
94.2%
56.7%
14.2%
13.2%
53.2%
28.4%
21.1%
0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%
Admissions
Alumni Topics
Campus Life
Community Relations
Courses
Events
Faculty
Financials
Institution, Overall
Inventions
Legal News
Other
Partnerships
Policies
Projects, Non - Research
Research, Agriculture
Research, Earth Sciences
Research, Life Sciences
Research, Other
Research, Physical Sciences
Research, Social Sciences
Staff
Students
Share of Engagement by Subject - ,External Blogs
Peer 1
Michigan State
Peer 2
Peer 3
Peer 4
The vast majority of discussion in external blogs is neutral.
Page 25
23
29
12
14
20
5
8
4
1
4
0
5
10
15
20
25
30
University of Michigan Purdue University Penn State Michigan State Arizona State
Share of Tone
Negative
Neutral
Positive
71%
3%
29%
94%
83%
42%
58%
6%
14%
58%
42%
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
Arizona State Michigan State Penn State Purdue University University of Michigan
Share of Engagement by Tone - External Blogs
Negative
Neutral
Positive
Emerging benchmarks Past PerformanceThink 3
PeerUnderdog nipping at your heelsStretch goal
Whatever keeps the C-suite up at night
Step 5: Define your benchmarks
26
27
Some benchmarks in social media Non-Profit EDU
% of Desirable Coverage 96% 98%
% of Key Message Communication
37% 2
Number of Messages tracked 11.5 5
Words per Key Message 14 7
Most frequent conversation types Express support (69%) Making an observation (28%)
Distributing Media ( 49% ) Giving a Heads up ( 19%)
% rallying support 3% 1%
% asking a question 4% 1%
% of exclusive mentions 17%
% mentioning brand in title 8%
% of discussions in Blogs 51% 38%
% of discussion on Twitter 30%
% of discussion on Facebook 8%
% of visibility from Social Bookmarking sites
9%
Overview of Key Metrics
Bookmark.
Ext. Blogs
Inst. Blogs YouTube MSM
SOV 2% — 8% 9% 11% 7%
Popularity
230 bkmks
500/mo. — 20 links
150k views —
Engagement 59 cmts 1 day 13 cmts
2-12 cmts 2 cmts —
% Positive 20% 32% 54% 50% 15% 15%
% Negative 0% 0% 4% 0% 1% 2%
Strat. Mess. 40%† 18%† 42% 42%† 18% 38%
Peer 1 was the competitive leader in all but YouTube, where Peer 4 and Peer 3 led.Actions attributed to individuals were responsible for most content, except on YouTube.
Rank Order
Facebook YouTube Social Bookmarking
External Blogs
Institutional Blogs
1 Campus Life
Events Courses Faculty Campus Life
2 Sports Campus Life
Projects, Non-Research
Research, Physical Sciences
Events
3 Technology Faculty Research, Physical Sciences
Institution Overall
Institution Overall
4 Product Services
Courses Events Expert Commentary
Institution Sub-Groups
5 Events Institution Overall
Faculty Events Admissions
Few subjects appear across all forms of social media, so tailor outreach accordingly
Step 6: Pick a tool
1. Content Analysis2. Survey3. Web Analytics
Step 6: Selecting a measurement tool
Objective KPI Tool
Increase inquiries, web traffic, recruitment
% increase in traffic#s of clickthrus or downloads
Web Analytics: Google Analytics, Omniture, Web trends
Increase awareness/preference
% of audience preferring your brand to the competition
Survey: Online -- SurveyMonkey, Zoomerang or Mail
Engage marketplace Conversation index greater than .8Rankings % increase in engagement
Web analytics or Content Anatlysis: TypePad, Technorati Omniture, Google Analytics
Communicate messages
% of articles containing key messagesTotal opportunities to see key messagesCost per opportunity to see key messages
Media content analysis –
% aware of or believing in key message
Survey
31
Content Analysis requires a content source:
Free: Google News/Google Blogs, RSS
feeds, Technorati, Social Mention, Twazzup,
$500+ Radian 6, Techrigy, Sysymos, Visible
Technologies, Scout Labs, Cyberalert, CustomScoop, e-Watch
32
A way to analyze content
AutomatedHuman:
Census vs random sampleSentiment vs TopicsThe 80/20 rule – Measure what matters because 20% of the content influences 80% of the decisions
33
A coding methodology
TonalityWhat messages were communicatedHow you’re positioned on key issuesDominance/Prominence/VisibilityAuthoritySubject of the article/postingWho was quoted?Products, events, initiatives, battles mentionedOptimal Content Score
What matters, what doesn’t
Standard classifications of discussion
• Acknowledging receipt of information
• Advertising something• Answering a question• Asking a question• Augmenting a previous
post• Calling for action• Disclosing personal
information• Distributing media• Expressing agreement• Expressing criticism• Expressing support• Expressing surprise• Giving a heads up
• Responding to criticism• Giving a shout-out• Making a joke• Making a suggestion• Making an observation• Offering a greeting• Offering an opinion• Putting out a wanted ad• Rallying support• Recruiting people• Showing dismay• Soliciting comments• Soliciting help• Starting a poll• Validating a position
Standard classifications of videos
AdvertisementAnimationDemonstrationEvent/PerformanceFictionFilmHome VideoInstructional VideoInterviewLecture
MontageMusic VideoNews BroadcastPromotional VideoSightseeing/TourSlideshowSpeechTelevision ShowVideo Log
Why an Optimal Content Score?
You decide what’s important:Benchmark against peers and/or competitorsTrack activities against OCS over time Positive:
Mentions of the brandKey messagesPositioningVisibility
Negative OmittedNegative toneNo key message
38
How to calculate Optimal ContentQuality score +1 0 -1
Score Score ScoreTonality Positive 3 Neutral 0 Negative -3
Positioning Contains 2 Doesn't contain 0
Positions the competition favorably or positions Sargento negatively -2
Messaging Contains 3 partially contains 0
Does not contain or miscommunicates key message (neg mess) -1
Quotes Contains 1 Does not contain -1Competitive mention
Does not mention Competition 1
Competition mentioned prominently -3
Total Score 10 0 -10
Visibility Score+1 0 -1
Score Score Score
Brand Photo Contains 3 Doesn't contain 0Contains competitive photo -5
Dominance Focal point 3 Not a focal point -1Visibility Headline mention 2 Top -20 % of story 0 Minor mention -2Target publication Top Tier 2 2nd tier 0 Not on target list -2
Total Score 10 0 -10
Optimal Content Score
Charting OCS over time between divisions
Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun Jul2007 2008
-100%
0%
100%
200%
300%
400%
500%
Optimum Content Score Relative to CompetitorsThe Percent Difference Between Each Business Unit's
Average Optimum Content Score and theAverage Optimum Content Score of Tracked Competi-
tors for each Business Unit
SAS
IDS
IIS
MS
% D
iffe
ren
ce
DIB
FBX-T
ALR-67(V)3
DDG-1000 MSE
APG-79, APG63
AESA for F-15E,Army MTS
VIIRS delays
Glory APS and VIIRS vs. com-petitors' EPX
ASAT, Patriot
ERGM cancellation
Patriot (Korea)European MD radar
RISS, GBS
MALD, AMRAAM
APG-63 (v3)ATFLIR
NPOESS; BOE B-52 jammer
RIS JPL
Trend against competition with OCS
Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun Jul2007 2008
0.00
0.10
0.20
0.30
0.40
0.50
0.60
0.70
0.80
0.90
Optimum Content Score by Company
Client
Company B
Company C
Company D
Company E
Company F
Surveys require:
A defined sampleA list – a way to get to that sampleAgreement on what questions you need to answerA survey instrument/questionnaire A testA way to analyze data SPSS SAS
42
Aspects of relationships
Control mutualityTrustSatisfactionCommitmentExchange relationshipCommunal relationship
43
Components of a Relationship IndexControl mutuality
In dealing with people like me, this organization has a tendency to throw its weight around. (Reversed)This organization really listens to what people like me have to say.
TrustThis organization can be relied on to keep its promises.This organization has the ability to accomplish what it says it will do.
SatisfactionGenerally speaking, I am pleased with the relationship this organization has established with people like me.Most people enjoy dealing with this organization.
CommitmentThere is a long-lasting bond between this organization and people like me.Compared to other organizations, I value my relationship with this organization more
Exchange relationshipEven though people like me have had a relationship with this organization for a long time; it still expects something in return whenever it offers us a favor.This organization will compromise with people like me when it knows that it will gain something.This organization takes care of people who are likely to reward the organization.
Communal relationshipThis organization is very concerned about the welfare of people like me.I I think that this organization succeeds by stepping on other people. (Reversed)
How to implement relationship metrics
Step 1: Conduct a benchmark relationship studyStep 2: Implement PR programStep 3: Conduct a follow up relationship studyStep 4: Look at what’s changed
Web Analytics Require:
Google Analytics/Web Trends/OmnitureUnique URLsData delivered in parallel with content analysisAbility to correlate and integrate data SPSS/SAS
46
Look for failures firstCheck to see what the competition is doing Then look for exceptional successCompare to last month, last quarter, 13-month averageFigure out what worked and what didn’t workMove resources from what isn’t working to what is
Step 7: Analysis - -Research without insight is just trivia
47
Ask for money Get Commitment Manage Timing Influence decisions Get Outside help Just Say No
Actionable Conclusions
48
Georgia Tech Case Study
Overall Comparison of Georgia Tech Social Media Outlets
50/17
0%
5%
10%
15%
20%
25%Share of Youtube
Share of Social Bookmarking
Share of FacebookShare of Institutional
Blogs
Share of External Blogs
Georgia Tech
Competitor Average
• Based on 2007 data, Georgia Tech outperformed its peers in Facebook presence, but significantly lagged peers on other social media.
• Post-2007 media monitoring has not included a social media dimension due to funding constraints, but this will be important to trend as feasible in the future.
Definitions: YouTube: a video sharing site. Social Bookmarking: a site where members can display media they have found on the web. Facebook: a social networking site. Institutional Blogs: blogs hosted and owned by schools studied. External Blog: any blog post that is not hosted by an institution.
Share of All Coverage
Where people get the content they share on Facebook
Sources of content
Genre of content
Understanding brand ownership of online video content
N=2,555,691
Peer Organizations
4.33%
Your Organization0.18% Other
Organizations8.65%
Individual Users86.84%
Use ownership to signal brand participation
Provide alerts for possible brand management issues
USO Case Study
54
Conversation Type January February March
Making an observation 142 152 46
Expressing support 20 40 261
Offering an opinion 18 1 1
Rallying support 4 4 5
Advertising something 10
Asking a question 1 2 6
Distributing media 4 2
Making a suggestion 2 3
Giving a shout-out 1 2 1
Calling for action 1 2
Expressing criticism 1 1
Disclosing personal information 1 1
Putting out a wanted ad 1
Expressing agreement 1
Recruiting people 1
Answering a question 1
Expressing surprise 1
Grand Total 206 213 321
0 50 100 150 200 250 300
Rallying support
Asking a question
Making an observation
Expressing support
4
5
22
40
24
221
Mentions
Conversation Type by Message Saturation
Contains no message
Contains One+ Message
0 50 100 150 200 250 300
Rallying support
Asking a question
Making an observation
Expressing support
12
178
34
83
Mentions
Conversation Type by Tone
Positive
Neutral
Negative
Moving conversation from observation to support
Social Media OverviewMarch 2009
56
SourceEngagement
Index
topix.net 19
community.babycenter.com 15
liveleak.com 12
patsfans.com 10
forum.militarysos.com, 9.7
michaelyon-online.com 4
forum.gon.com 2
americamatters.org 1
outboundnews.com 1
flickr.com 0.7
us.allscoop.net 0.67
moaablogs.org 0.5
swymer.wordpress.com 0.5
youtube.com 0.1
0 5 10 15 20 25
icelebz.com
us.allscoop.net
buzzhollywood.com
topix.net
video.aol.com
forum.militarysos.com,
flickr.com
youtube.com
1
2
1
2
5
3
2
1
3
5
4
21
16
Top Sources by Message Saturation
Contains no message Contains One+ Message
Media Engagement & Online Giving
Red line indicates media impressions
35,152,789 OTS
6,253,852 OTS
Best Practices:
Correlations to bottom-line impact
DonationsMembershipsSign-upsLeads
Using SMM for planning
Define the time frame, market/topic you want to studyUse Google News, Technorati or Radian6 to identify the conversations around the topic Analyze the conversations for type, tone and positioningLook at share of positioning, tone or conversation
Benchmarking against your peers
Looking at what the best doSetting goals accordinglyUse data to persuade recalcitrant spokespeople
Social Media in CrisisListen instantly to a wide range of influencersIdentify weaknesses in communications, customer service, or in the product
Improve your reputation
Listen first, then respondStop doing stupid things
Thank You!
For more information on measurement, read my blog: http://kdpaine.blogs.com or subscribe to The Measurement Standard:
www.themeasurementstandard.comFor a copy of this presentation
go to: http://www.kdpaine.comFollow me on Twitter: KDPaineFriend me on Facebook: Katie
Paine Or call me at 1-603-868-1550
