Perficient Digital Transformation Blog


Posts Tagged ‘social’

Lessons from 2014: How to get more clicks on Facebook.

Merry Christmas!  As I have some time off at the end of the year, I’m looking back at information I have gathered in my reading list that I find interesting.  I came across the article We tested all the best advice to get more clicks on Facebook. Here’s what worked by Kevan Lee at Buffer’s Social blog.  As I re-read the article I had a funny feeling that I’d commented before on posts by Kevan Lee.  Sure enough, I found two other blogs posts from this year that contained information from Kevan.

What is intriguing to me about this article is that Buffer used a very methodical approach to testing each of their theories.  Too often I see companies just try things without really following a good scientific methodology.

First Buffer started with a baseline of how their Facebook page performed.  This is critical because you can’t measure what works without having a baseline.

Here are the seven techniques Kevan used to see which were the best at getting users to click on a Facebook post:

  1. Post to Facebook at non-peak times
  2. Post more frequently to Facebook – six times per day
  3. Post less frequently to Facebook – once per day
  4. Ask questions in the updates
  5. Change the style of the update
  6. Post only link updates
  7. Post different types of images with the links

That seems like a pretty good list of techniques. I won’t go through the results for each test here – you can read through Kevan’s blog post for the details.  However, here is an example of the results from the first test – posting at non peak times:


Here they found a big increase in clicks at 11:00 pm, even though they were posting less frequently at this time.  Very interesting.

Here are the three techniques that worked the best for Buffer:

  • Share link posts
  • Share in the evenings
  • Create a main image/graphic for your post

Of course this is data only for Facebook for Buffer.  You should follow a similar test regime to see what works best for you on Facebook and other social media sites.

Lessons from 2014: The Problem with Sentiment Analysis

As we wind down 2014, I’m taking a look back at some items in my reading list and bringing forward the ones I found important from a learning standpoint.  The article The Problem with Sentiment Analysis by Sarah Kessler at Fast Company in November 2014 qualifies as one of those “aha” articles.

Analyzing social media has been a hot topic in the past couple of years.  Ms. Keller points out that during the 2012 presidential election season USA Today had a daily story about President Obama’s “sentiment” score versus Mitt Romney’s score.  The score was calculated by analyzing social media posts about each candidate.  In theory, the analysis could show which candidate is getting more positive comments versus negative comments.  And, in theory again, this could tell us about public opinion for each candidate.

However, Ms. Keller interviewed Marc Smith who pointed out that this type of sentiment analysis is inherently flawed.  Marc Smith is a sociologist who specializes in the social organization of online communities.  He went so far as to say about the USA Today stories that “This is remarkably poor data. That this is borderline criminal.”  As Keenan Thompson says in Saturday Night Live, “What up with that?”

Tight Crowd Pattern Graph: NodeXL Graph Gallery

Mr. Smith argues that this type of sentiment analysis only reveals which group of supporters “shouted” the loudest that day. It really tells us nothing about public opinion of either candidate.

What I found really interesting is Mr. Smith’s work on how crowds form around a topic on social media in six different shapes:

  • Polarized crowd in which two groups form and rarely interact with each other.  This is akin to the political sentiment described above.  The Obama crowd did their thing and the Romney crowd did theirs independently.
  • Tight crowd where a small group of people cluster around a conference.  The image in this post shows a tight crowd pattern.
  • Brand cluster in which people talk about a brand, but rarely interact with each other in the crowd
  • Community cluster where multiple small groups form
  • Broadcast network in which many people connect with a media outlet, but not with each other
  • Support network where something like a service center connects with lots of people, but those people don’t really interact with each other.

According to Mr. Smith, looking at the shape of the network lets you see that not all social media posts can be treated the same.  He argues that you should report on the size, volume and content of each major cluster over time.

This makes a lot of sense to me.  If you are measuring sentiment for a particular brand, event or anything else, you should be interested in how the network aligns with your goals.  Say you are hosting a conference – you’d want to see the social network form into a “tight crowd” pattern.  If it doesn’t then something is probably not working right.  Likewise if you are a brand and see the “brand cluster” pattern emerge, you may want to take steps to encourage your followers to interact more with each other, if that is a goal.

If you are involved in sentiment analysis or any social media analysis, I encourage you to follow the links here and take a look at Marc Smith’s research in this area.  Let me know what you think.

G2 Crowd scores Salesforce Chatter top social collaboration tool

I saw the following tweet from Salesforce CEO Marc Benioff:

Image of tweet from Marc Benioff

This was interesting enough for me to follow the link to see why Salesforce Chatter was rated the best social collaboration system.  Barry Levine at Venture Beat had a nice article about G2 Crowd’s grid on social collaboration.  Here are the first two paragraphs:

Salesforce’s Chatter is the social collaboration tool with the highest customer satisfaction.

That’s a key takeaway from G2 Crowd’s new Grid for Social Collaboration tools, the first Grid report for this category. It scored 14 business products based on G2 Crowd’s customary metrics of customer satisfaction and business presence, in this case based on over 325 reviews from business professionals.

This was really interesting to me.  G2 Crowd allows people to rate these systems and then translates those ratings into a 2×2 grid similar to Gartner’s Magic Quadrant.  Venture Beat also mentions the 325 reviews used to generate the ratings.

While the G2 Crowd rating system looks good and in theory produces crowd-based results, when you dig a little deeper, you find that G2 ratings are really based on much smaller sample sizes. For example, here is the latest grid as of today:

G2-Social Collaboration

In looking at the Leaders Quadrant, you find that Chatter had 97 reviews and Microsoft Lync had only 11 reviews.  You can click on an icon to see the number of reviews:


I don’t dispute that Chatter is an excellent social collaboration tool and Salesforce should be proud to share these kinds of results.

However its hard to rank these tools on customer satisfaction alone from such small numbers of reviews. First, I imagine you could easily get 11 reviewers to enter lower scores for any product. Likewise, I’m sure you can ask 11 people to go rate a product very high.  I’m not suggesting anybody is doing this on the G2 Crowd site to impact the rankings, its just a possibility.

Second, and more importantly, assume that the total population of users for all these tools were 1,000,000 users.  That’s probably very low, but its a nice round number.  To have any confidence in our survey, we’d need a sample size of over 1,000 users responding to the G2 Crowd reviews. And that would be 1,000 random users.  With only 11 reviews or only 97 reviews, statistics tell us that we should not be very confident in the results.

The G2 Crowd information is interesting and it would be great if we could get the right number of reviewers to make the data statistically meaningful.  But, really, don’t rely on this information alone to make your purchasing decisions.  That’s why Gartner, Forrester and Perficient use a variety of factors when rating these systems.



A Day in the Life of a Social Media Manager

In the BufferSocial blog, Kevan Lee posted an article for Social Media Managers.  The post takes a look a “typical” social media manager’s day and breaks down that day into many different activities, represented in the info graphic here.   

Mr. Lee also provides several different views on how other people spend their days managing social media.  One person, Finola Howard, manages to compress all her daily activities into just one hour per day. Her tasks include:

  • Use SocialOomph to figure out which new twitter followers to accept
  • Measure which posts are performing the best so you can take advantage of them
  • Schedule tweets and posts for the day.  She uses Buffer for this, other use tools like Hootsuite.
  • Find content
  • Respond to others
  • Monitor engagement of fans and followers

In general, the post identifies 12 tasks of a social media manager.  The twelve tasks are shown here and the article does a great job of explaining each of them.

If you manage social networking within your company, say using IBM Connections, Yammer, Jive or others, you should also pay attention to the tasks.  Each of these 12 tasks apply to internal as well as external social managers.

In addition, Mr. Lee provides a series of checklists for the social media manager.  These lists come from places like Mindbrew Creative, HeroX, Hootsuite and others.  Even by itself, the various checklists are well worth your time to understand.

Overall, A day in the Life of a Social Media Manager is extremely valuable and full of great information.

Why Social Needs to be Part of Your Portal

IBM’s 2014 Digital Experience Conference started Wednesday off with a great session by Mac Guidera, Social Workforce Strategist from IBM, titled “Why Social Needs to be Part of Your Portal.”  The session was very insightful blending a mix of statistics, trends, best practices and insightful thoughts.

Why Social Needs to be Part of Your PortalSocial Business Patterns

Patterns represent modernized processes with dynamic, repeatable and measurable “people interactions” created by building social into work and life.  These patterns are repeatable way to interact an engage, share innovative ideas, finding out who knows what and find information.  Key patterns include:

  • Customer Engagement
  • Marketing
  • Sales
  • Customer Service
  • Recruiting
  • Supply Chain
  • Expertise and Knowledge
  • Workplace & Safety

Each of these patterns has value propositions and ROI, for example Customer Service maps to customer satisfaction, increased revenue and efficiency.  Marketing can map to awareness, marketing effectiveness and trust. Read the rest of this post »

Cool Tools by Gartner

Gartner analysts, Tom Austin and Mike Gotta, presented some cool tools, that have a social angle, primarily Virtual Personal Assistants (VPA)

Welltok – a virtual personal assistant for health care, see CafeWell as an initial application

  • – community for health and wellness
  • – provides advice to individuals

MindMeld – conversational application, connect through Facebook, conversational assistant – listens, finds, organizes, archives.

– advice to groups of people

GridSpace – meeting memo taker – preserve and analyze meetings

– who said what to whom in a meeting

HP in Context Analytics – auto-match searching based on text chat in MS Lync, later other IM’s and voice.

Wrike – as work is less routine, people self-organize work – realtime project co-ordination and work management platform. Social Task Management

Microsoft Office Graph – social queues and behaviors – precursor to virtual personal assistant, in Office 365 only.

  • – includes app and workload intelligence
  • – auto population of Office Graph
  • – Big Data analytics
  • – Semantic expressiveness
  • – Graph based search

Lots of power inside Office environment. Less so for departments less dependent on MS Office, i.e. Engineering with CAD, or Marketing with Adobe.

Sqwiggle – continuous spresence awareness across remote teams – making remote workers feel like they be log.

Highspot – influence, authority, relationships – manually tag information, search, and discover. Who do you follow, who follows you.  Highspot is here today, and it works, and is available in the cloud.

Cuff – personal security device that fits into an expanding line of jewelry. Pairs with you iPhone or Android phone, and with one press will alert loved ones when you need them. You can also send reminders, or notifications that you are trying to reach the person.

Interesting papers that may be worth a read:

  • Cool Vendors in Social Software and Collaboration. 2014 – Mike Gotta – G00262576.
  • Cool Vendors in Smart Machines, 2014 – Tom Austin – G00262488.

The rise of smart machines … the dawn of automated workers

Tom Austin presented on the rise of smart machines and their application in the work place, and society at large. I think this is pretty cool technology, and will revolutionize the way man and machine interact over the coming decades. It will probably come faster than we keep up with and Uncle Sam will probably get in the way. So no self-driving cars, or fleets of trucks, or even freight trains, any time soon. Even though these would help reduce CO2 emissions and make roads safer.

Smart machines are not general purpose computing – they are not really smart – they operate in a specific environment (i.e. volvo vs airbus).
Smart machines may scare us – under what conditions will we trust technology to make decisions that we used to make?

  • they operate autonomously
  • appear to understand abstract concepts

Example of machine learning: Google deep neural network – analyzed 10 million frames from random youtube videos and was able to identify 20,000 classes of images in 72 hours – detected cat faces, two people kissing, dancing – with a very high degree of accuracy.

The algorithms based on 2005 – 2008

Microsoft demo – Rick Rashid spoke English, yet audience heard Mandarin

  • Deep neural nets are one key to learning and understanding
  • images
  • faces
  • emotions, etc
  • large bodies of unstructured content

Autonomous movers – robot in amazon warehouse, Caterpillar truck (autonomous driving), google cars (self driving) – autonomous cars on streets of gothenburg,
x47 drone

Sages – virtual personal assistances (focus on context)
Apple Knowledge Navigator (you tube video)
Apple Newton
Siri (precursor to smart personal assistance)
Search: Eric Jorvtix TEDx 2013”
Google Now:
Knowledge Graph
Deep Neural Network
End of search

Smart Advisors (focus on content)
IBM Watson
not a personal assistant
deep but narrow knowledge of content
Watson app (Memorial Sloan Kettering)
clinical oncology
breast cancer treatment recommendations
advanced natural language techniques
Natural language generation
fool the reader or listener
narrative science / yseop (easy-op)

Cornell Robot – predictive physical assistant
People and technology working together
Laggards lose
collaborate with machines you trust
replace people
PEDs – performance enhancing devices
47% of US jobs are at risk over the next decade or two – Frey & Osborn, Oxford, September 2013

34% of careers will be enhanced by smart machines

IT matters a lot
BYO smart machines will thrive
privacy, security and innovation are at odds
single vendor will be at a major strategic disadvantage

Smart Machines
engage, empower and delight employers
drive while intoxicated

Smart Advisors First

  • should you exploit a smart advisor
  • should you buy one?
  • how can these make the highest paid workers perform better
  • Virtual Personal Assistance
  • Leverage information explosion
  • Many assistants coming
  • Users will discover by doing if they can
  • Innovation, privacy, and security are at odds


  • Consider consumer-grade business opportunities now
  • 2017 will be the year these technologies take off at work
  • Your users need to start experimenting heavily by 2015


Non-routine jobs are increasing
Routine jobs are in decline (60% in 1976 – 40% 2014)

New “Worksplace” Strategy,
man-machine collaboration
helping you excel at difficult tasks


  • Get Smart
  • calls for IT leadership, not just management
  • Engage the business
  • Respect the impact on people
  • impact of software and robots on employment, work, and careers of people will be profound
  • Read: “Cool Vendors in Smart Machines in 2014” – Gartner Paper

Case Study: Implementing Social-Based Collaboration

At the Gartner Portal conference, John Stepper, Managing Director at Deutsche Bank told his story of implementing a social network within a large German bank.

Most (large) companies are stuck in finding the right people and data

  • Early AdoptersEarly adopters – usually the same people
  • Little lasting change occurred
  • Introduced MyDB – Social Network – basically rebranded Jive OOTB
    • Modern Backplane – more appropriate than meetings and email in many circumstances
    • This pilot stuck and made a difference
    • Adoption is increasing
  • 7 Elements of an enterprise collaboration strategy
    • Platform
    • Commercial value community managers
    • management engagement
    • advocate network
    • center of excellence
    • individual benefits
  • Application to own organization is more important
  • Success looks different at different levels
    • what worked 2 years ago may not work now
  • 7 Questions
    • Can we do that? (What have our proposal, and business case, but there are plenty of people that are ready to say no! HR, Legal, Workers council)
      • Yes, we can.
      • Platform is opt-in. People chose to use the tool.
      • Ratings and private groups where turned off.
    • What if people do something stupid – people say something they shouldn’t?
      • e.g. @queendemetriax_ tweeted she was an islamic terrorist getting ready for a big event against American Airlines. She was a 14y/o girl and was arrested by the FBI.
      • No anonymity – Good policies – 1-click to flag content (rarely used)
    • What’s it worth?
      • Initial business case based on intranet spent – well under the true value
      • 50% emails
      • 30% meetings
      • learned helplessness
    • Will anyone use it?
      • Got budget for 12,000 people to use it once.
    • Is it official?
      • Is it for the unwashed? IT? Backoffice?
      • Is it for real work?
      • People thought it was “Facebook” – and said I’m not using that.
      • People wanted something official, with prestige. – Barak Obama.
    • Will businesses use it?
      • Content and convenience are the killer apps
    • What else can we do with it?
      • Other 6 dragons have been slain.
      • We can make work more fulfilling
      • People hate work – it’s dehumanizing


  • Working out loud
    • Bryce (Eli Lilly)
    • Using social platforms at work: something like OPENPediatrics.
  • “Without myDB, our project would be much more difficult”
  • “I’m happier”

Get ready for the Digital Workplace

This week I’m at the Gartner Portals, Content & Collaboration Summit, in Los Angeles. Jeffrey Mann and Susan Landry of Gartner delivered the opening keynote: DELIVER EXCELLENT ENGAGEMENT THROUGH MOBILE, SOCIAL AND ANALYTICS “The prowess with which you engage your customers, employees, partners and constituents is the single most important determinant of your organization’s success. Yet it remains among the most challenged of activities for most organizations. Advances in social media, content analytics and ubiquitous mobility are already smoothing the path toward more effective engagement. The future holds even further hope. We’ll zero in on the future tactics and technologies you’ll need to move beyond failing or, at best, ho-hum engagement, and toward an era of truly excellent engagement.” 4 key themes where addressed 1) Engagement 2) Digital 3) Content 4) Integration 1) Engagement – becoming more engaged with your workforce, customers, constituents. Only 13% of global workforce felt engaged. 2) Digital Is really about enabling the “Digital Workplace”

  • Which enables new and more effective ways of working
  • improves employee engagement
  • exploits

Example: Story: see

  • Story is a retail story in New York, that uses digital heat maps to gauge what products customers are interested in, and know how to position things.
  • It reinvents itself every 6 weeks around a theme like a magazine does

3) Content

  • Content Management – disliked term – mitigating a problem
  • What do most applications think “Content” refers to:
    • Documents
  • What does content actually mean?
    • stuff – that passes desks, minds: files, documents, movies, images, tweets, etc..
  • Not just what your business does … but what it thinks.
  • Why does it matter?
    • connects the dots and connects the thoughts
  • Why does it matter to your business
    • it’s the grease between decisions
  • Content is the oil of the 21st Century
  • can be exploited (like oil)
  • Get the right content and the right time

4) Integration

  • back in vogue
  • what is new?
    • IT wants to connect two or more applications to share data – like SOA – not new
    • digital work place needs more
    • bring about the right experience for the user
    • many more sources of content
    • integration shields users from complexity
  • What does integration look like
    • Dr Jeff Burns is the Chief of Critical Care Pediatrics at Boston Children’s Hospital
    • Worldwide there is a shortage of trained doctors
    • Open Pediatrics goes some way to addressing this by connecting doctors to patients wherever they may be.
    • Example of Digital Workplace
    • Extensive use of video
    • audio is transcribed to text
    • creative – animation
    • audio enhancements – to ensure it’s crisp and clear

Example of a Digital Work place 2): System to automatically know what is needed based on context – scenario involves a sale rep going out to make a presentation to a client

  • traffic updates
    • automatically updates alarm clock to wake up earlier to account for traffic / road works
    • updates GPS with best / fastest route
  • news updates, about the company you are meeting with
  • sales documents
  • contracts
  • real-time product / inventory updates and alerts
  • capture and transcribe results

Recommendation Portal

  • Images Documents Likes Upsell
  • Reviews Transactions Data Cross-sell
  • Maps Comments Orders Actions

Government Finance HR Take the lead to create a digital work place: What Do I Do Now?

  1. Create and imagine a scenario
    1. Develop three scenarios for your organization of what could be.
  2. Identify parts of your infrastructure that must go (which parts are hostile)
  3. Identify the skills you can afford to lose, and those you need to develop?

Marketecting the enterprise?

At the Association of Enterprise Architecture Summit in Austin, Texas last week, John Zachman was the speaker of honor. For those unfamiliar with his work he is the leading proponent of Enterprise Architecture, and I don’t mean that in the marketing sense where all companies are the leading in “blah blah blah”. As a career IBM’er, he is considered by many to be the founder of the modern craft with The Zachman Framework for Enterprise Architecture, although he refers to it as an The Enterprise Ontology.

The emergence of a social enterprise marketecureZachman reminded us of how controversial Nick Carr’s seminal article “IT Doesn’t Matter” was 10 years ago, when most of us thought IT would by itself revolutionize business and the world in general – this was in the aftermath of the “dot.gone” era. He then went on to say that Enterprise Architecture, or EA, is not a technology issue, but rather an enterprise one, and that the role of the EA does not really belong in IT. This was confirmed later by some of the other guest speakers.

To paraphrase Zachman, ‘Over the last 75 years or so, people, or more accurately the roles fulfilled by people, have been systematized and automated. These systems essentially represent the enterprise as a whole. An EA possess the engineering skills to design artifacts used to engineer an enterprise’.

Zachman also called on the work of Alvin Toffler, of Future Shock fame – which by the way is still amazingly relevant, perhaps only more so, not in it’s specificity, but more in it’s approach as to how much change has been going on in the world and how we struggle to adapt to it.

Talking of customer expectations, he explained that all customers expect a custom experience from an enterprise. They want a custom enterprise. He threw out the a challenge to all willing to accept it, how will your enterprise become a custom enterprise?

At this point, I got thinking about my area of expertise, namely Portals, Social, and Web Content Management technologies. In other words digital experience technologies. The digital world certainly can provide very large organizations the means with which to provide custom products or services to customers. Remember custom Nike shoes? Or Dell computers of a decade or so ago? Where these early examples of custom enterprises? Digital experience technologies empower enterprises to provide a custom experience tailored to exactly the needs or desires of a single customer, and at relatively low cost.

Zachman provided a definition of architecture by means of several colorful examples, “Seven thousand years of history suggest the only known strategy for addressing complexity and change is architecture.” Think of the hand axe, throwing stick, or shaduf, all examples of architecture, in some form, at work, in that each design or blueprint that may be used by craftsmen to build from or improve upon. He gave the example of the Coliseum in Rome. This is a static building, and not architecture. Architecture was the process of planning ahead of time. It is the set of descriptive representations relevant for describing complex objects.

As it it with modern digital experience platforms. The implemented platform is not the architecture. The architecture is the process of planning the implementation ahead of time. It includes understanding the business need and outcome, envisioning how the modified business will operate, determining how to reach the desired state, and as well as understanding how any new components or processes will fit in with existing ones. In other words, implementing a digital experience platform involves a lot more work than only selecting and configuring the technology. It involves a significant amount of planning ahead of time, or upfront enterprise-wide architecture.

Mike Walker, the president of the Texas Chapter of the AEA, made a few other interesting points regarding the EA profession in general, that I believe are also relevant to large scale technology initiatives such as transforming an enterprise through digital experience platforms. Many people involved in, or doing, Enterprise Architecture, today come from an engineering or technical background, they often have high IQ’s and are great at explaining the “speeds and feeds” of a set of technologies. They are also often found reporting to the CIO. Psychologist have found that people don’t make decisions based on what the neocortex is telling them (data), but rather the limbic system (emotions). Something that I experienced first-hand over the weekend, whilst looking for somewhere to live, I had all the data that said that staying in Austin makes sense: lower overall taxes, lower rents, live music, etc, compared to moving to California, higher overall taxes, higher rents, ocean. Usually the move is the other way around, however as a surfer and sailor, my limbic system won out over my neocortex. And that brings me back to Mike Walker’s point, engineering, or solution architecture, is often performed by introverts. Enterprise Architecture requires socialization across an enterprise to make it successful.

I argue that enterprise architecture is not really a role, but rather a practice, perhaps within a wider center of excellence. An EA practice would be made up of a broad range of complementary abilities and skillets which can only enhance it’s value. This is perhaps where internal marketing can help. Marketing people tend to be extroverts, and more attune to getting a message out and understood. This may lead to a simplification of the more detailed enterprise architecture, and the emergence of a social enterprise “marketecture”, but if that aligns the stakeholders and makes for a successfully adopted system, that’s all the better.