21. 21
In the world of NPR…
Broadcast radio listening
Digital analytics
Donations and membership
Underwriting revenue
Important things to measure, but do they tell the whole
story of what you are trying to do and be?
Are there things you’re doing less of simply because they’re
not as easy to measure?
We tend to measure what is familiar and
easy to count
22. 22
Community impact
More informed public, influence on culture, policy
change, etc.
Audience insights
Listener attitudes, unmet needs, changing behaviors
Digital
Less familiar to many people in our organizations
Constantly changing and hard to keep up with
Overwhelming amount of data at our fingertips
It’s harder to measure what is less quantifiable,
newer, changing, or difficult to track
46. 46
Yes, this sounds like Lean Startup thinking
Assumption/hypoth
esis
Minimum Viable
Product
Analytics,
research, testing
47. 47
Designing choice in the new NPR One app
Original design: Give
users more
control/choices
Hypothesis to test:
Fewer immediate
choices + simplicity
= longer listening
Results: Listening
time up 8%
63. Putting Data to Work
How Digital Analytics Informs Operations
Analytics Academy
March 19, 2015
64. Overview
What is your business goal?
How does digital communications
support that goal?
What are the goals of digital strategy?
What can you learn about user behavior to
determine if you are achieving that goal?
How can you use that knowledge to improve
your strategy?
68. Goals of the Website
• Help prospective applicants discover the
Harvard experience
• Hear authentic current student voices
• Understand the application process and
successfully complete the application
• Access information about application
process and admission decision
70. Interpreting the Metrics
• Seasonality - The Admissions cycle affects traffic to the
site. By understanding that traffic we can plan for server
capacity and uptime during our most critical time.
Regular Decision Early Decision
71. Understand Traffic Sources
• Traffic Acquisition - How does it change
over time?
January 1, 2014
January 1, 2015
72. Analytics-driven Content
• Landing pages: identify
where users are
entering your site, other
than the homepage
• Highlight key messages
• Make content easily
accessible across the
site
73. Understand Traffic Sources
• Mobile Traffic - Mobile traffic has grown from 15% in
2014 to 20.5% of all traffic in one year on smartphones
alone. On EA decision day, mobile traffic was nearly one
quarter of all traffic.
74. Mobile Strategy
• Optimize for mobile.
Assume users will view
on mobile first
–Responsive design. All
content must work on
mobile– web, email,
tour
–Test early and often
on many devices,
browsers
75. Social Media & Google
Analytics
Acquisition > Social > Users Flow
76. Social Media
Many sources, constant change
• What can we measure?
We can’t we measure?
• Aggregating social data
from various platforms
• Measuring engagement at
the post level to inform
content strategy
78. Informing Social Media Content
• Our audience are 16-18 years old.
• We are monitoring usage to go where they
are, rather than asking them to come to
us.
79. EMAIL & OUR SEARCH LIST
Strategies for outbound communication
80. Goals of Outreach
• Search list are highest priority users–
actively recruiting these prospective
applicants
• Targeted outreach focused on students
from diverse socioeconomic and cultural
backgrounds
• Drive traffic to our website, engage with
our students, start an application
81. Measuring Email Effectiveness
• Engagement is measured in
open rates, clickthroughs
• Deliverability, bounces, quality
of list
• Use UTM codes to link back
to website
• Send through Slate,
integrated with admissions
database. More on that later!
82. Google URL Builder
● utm_source: Identify the advertiser, site, publication, etc. that is
sending traffic to your property, for example: google, citysearch,
newsletter4, billboard.
● utm_medium: The advertising or marketing medium, for example: cpc,
banner, email newsletter.
● utm_campaign: The individual campaign name, slogan, promo code,
etc. for a product.
● utm_term: Identify paid search keywords.
85. Email Content
Optimize based on learning
• 70% of emails opened on mobile. Design
for mobile
– Responsive design
– Length of message
– Placement of links
• Know when to send based on audience –
3-5pm for our audience
• Test subject lines
– A/B testing
87. CRM & Admissions Database
• Content strategy is based in goals
for recruitment & yield. CRM tracks
progress towards that goal.
• Also makes online reading of
applications possible
• Implemented this year- opportunity
to reevaluate strategies and better
analyze our data!
88. Student’s Path in App Cycle
• Applicant data is centralized throughout all
stages of the process– as a prospect,
inquiry, and applicant.
Prospect Inquiry Applicant
89. Getting Specific with our
Message
Recruitment communications for targeted
prospective students
Low socio-economic
status
seniors
Hispanic, African
American, Asian
American, Native
American
Low Application
schools
Low SES Juniors
HFAI UMRP HCC
First in the
family to graduate
from a four-year
college
First
Gen
90. Measuring Digital Recruitment
Track offline interactions
and aggregate data in
Slate
• Interactions with
current students
• History of engagement
is kept in the applicant
record
• Query and pull reports
based on interactions
97. What does ‘success’ mean?
● Metrics paired to business goals
o Pageviews (reach)
o Time on site (attention)
o Email signups (loyalty, audience building)
● Anecdotal feedback
o internal (local leadership, faculty)
o external (media, influencers, awards)
98. What does ‘failure’ mean?
● Failed to meet expectations
● Elicited negative responses
● Opportunity cost
101. Campaign checklist
Using reporting from start to finish
1. Decide what’s worth sharing
2. Build consistent tracking methodology
3. Share prepared content
4. Implement analytics collection process
5. Put analytics through feedback loop
6. Evaluate effectiveness, determine
takeaways
102. Deciding what to share — and how to share it
Editorial decisions
*Editorial calendar
*Trending topics (Twitter, FB, Iconosquare)
Google trends
Topsy
Encore Alert and similar services
Format
Platform updates
Mobile proliferation
*Current research (Sharing Fast and Slow)
104. Choose content: Format
Sharing Fast and Slow
104
POST PRIMED FOR SHARING/REACH POST PRIMED FOR INTEREST/CLICKS
200 clicks to story176 retweets, 13 new followers
150. -Assistant Director of Alumni Marketing and
Communications at the Harvard Business
School where I manage the office’s
social media channels
-Have helped generate over 18,000 unique
alumni interactions via Twitter since January
2013
-Graduate of UMass-Amherst and the Tufts
University Graduate School of Arts and
Sciences
-Author, Social Media Matters blog,
http://robertbochnak.wordpress.com/
-Superhero sidekick of Aniceta
15
0
A little bit about me….
152. Our Approach: Identify HBS alumni who are on
Twitter and add them to the appropriate list.
15
2
153. Our Approach: Engage alumni through retweets, original
tweets promoting their work, tweets sharing events, tweets
connecting alumni with each other, tweets to alumni based
on their interests, etc.
15
3
154. Our Approach: Leverage the relationships we have
developed with alumni so they participate in discussions.
15
4
155. But before all this engagement could happen we
needed to collect data on our alumni….
So we decided that each alumni interaction we had
on Twitter would be tracked on an excel
spreadsheet. This sheet would include information
drawn from the Twitter profile and alumni directory
file of each alumnus/a and would include
information such as city and state of residence,
interest, year of graduation, and section.
15
5
157. Data Use
With this data—as earlier slides conveyed—we’re
able to send targeted tweets to alumni based on
their interests.
15
7
158. Reporting and disseminating the data
-Each interaction with an alumnus/a is “scored” on
the tracking spreadsheet: 1-2 “touches” a month
equals “minimally engaged”; 2-4 “touches” a
month equals “moderately engaged”; and 5 or
more “touches” equals “highly engaged.”
-Interactions are retweets, favorited tweets, or
original tweets by individual alumni.
15
8
170. BEWARE OF METRICS BLOAT
○ Numbers replace your business goals
○ Confusion among team members
○ Lost alignment within organization
○ Spread thin aggregating metrics
○ Metrics no longer used to inform business
decisions
171. HOW DO YOU TREAT
METRICS BLOAT?
a. Invest in more fancy tools
b. Wither and die
c. Other
191. ASK QUESTIONS
○ What happened in February?
○ Was the creative different?
○ Was the email sent at a different time?
○ Did the email go to the wrong list?
192. CHOOSE YOUR ADVENTURE
○ Wait for more ‘beacon’ data
○ Tweak your strategy (one variable at a
time)
○ Follow the trail with more analytics
199. CATHERINE’S EMAIL
METRICS DETOX PLAN
1. Identify your
business goals.
2. Connect your
communications
strategy to your
goals.
3. Track the end goal.
4. Work backwards. 5. Tweak
233. London Underground Tube Map
▣1908
Underground lines were
laid out geographically,
often superimposed over
the roadway of a city map.
Source: datavizblog.com/
236. London Underground Tube Map
▣1933
Beck believed that
passengers riding the
trains were not too
bothered about the
geographical accuracy,
but were more interested
in how to get from one
station to another, and
where to change.
Source: datavizblog.com/
241. ‘’
From 2013 to 2020, the
digital universe will grow by
a factor of 10 – from 4.4
trillion gigabytes to 44
trillion. It more than
doubles every two years.
Source: emc.com
242. ‘’
Most companies estimate
that they are analyzing a
mere 12% of the data they
have, according to a recent
study.
Source: Forrester Research
259. Daily Gazette report
What: Quick view of daily Gazette
email performance
Who: Senior leaders, Gazette
editors, Media Relations, Digital
Strategy
Why: Identify what resonates with
readers, engagement trends
Place your screenshot here
262. Weekly Gazette web
traffic
What: Google Analytics dashboard
Who: Senior leaders, Gazette,
Media Relations, Digital Strategy
Why: Identify overall traffic, explain
sources of traffic, anomalies, review
owned traffic performance, social
impact
263. Weekly Gazette web traffic
Identifies not only top overall stories, but
highlights traffic generated from owned
sources like Harvard.edu, Daily Gazette
email, and Harvard’s social channels.
264. Weekly Gazette web traffic
▣ A written summary of select
referral traffic, including
influencers on social media
▣ Identifies media pickup and
impact on traffic
265. Monthly report
What: Combine metrics across
various digital platforms
Who: Senior leaders, HPAC, AA&D
Why: Analyze trends, large spikes
and dips, measure against previous
month/year
267. Monthly report
▣ Provide most important findings and with each
report; explain changes and future opportunities
268. How to get started
▣ Define the goal and the audience
▣ Identify metrics that measure the goal
▣ Collect metrics consistently
▣ Simplify to show performance against the goal
▣ Present insights that are memorable and lead to
action
YOU control the context.