Three Principles for Measuring the Value of Enterprise Social Networks

OLYMPUS DIGITAL CAMERAI had the opportunity to speak at Webtrends Engage last week about social media measurement, both external and internal. On the external social media side, we’ve been wrestling with the Great Philosophical Questions for some time:

What is the ROI? What does engagement mean? What is the value of a like or fan? But it’s a different story when we apply these questions to an internal organization.

Measuring the impact of social collaboration within an organization is different from measuring social media throughout an external community. 

When you look inside your own organization, you’re hoping to see whether collaboration is having a detectable impact on real business priorities: customer satisfaction or NPS scores, revenue, brand sentiment, employee retention, operational costs. Of course, within organizations, as in the external world, there are many objective and subjective factors that you have to take into account–salaries, company success, technologies, training, and many more–that influence your findings.

But to make and justify an investment in social collaboration technology, you have to start somewhere; the hypothesis (better collaboration should lead to happier employees and customers, and therefore untold riches) will take you only so far.  But all social media is not created equal. Following are three core principles to keep in mind when measuring the business impact of social collaboration.

#1: The Exchange of Value is Reversed 

When we interact with a brand, we are free, independent, and hold all the cards. We may be influenced by our peers, but we don’t have to be. It’s our choice. And we are the buyer; of goods, services, information. But when we interact with an employer, different dynamics come into play. We are, let’s be honest, in a different kind of power relationship. In many states, we are at-will employees. The exchange of value is reversed. We are the goods, and the employer is the buyer. So we don’t have the same independence that we do as consumers.

Point #2: Identity is Known and Constant

On the Internet, your identity is fluid. You can use your real name on Facebook, a handle on Twitter and a different handle altogether on Instagram, Tumblr and Pinterest. It’s challenging for companies to determine who you are, because there is no universal social login. Identity can be as transparent as a first-initial-last-name Twitter handle (@setlinger), as confusing as in the Manti T’eo affair, or completely anonymous. Companies like Janrain, which provides user management and social login services, have made a business out of helping businesses better understand user identity. But in an organization, your identity is known and constant. There’s nowhere to hide.

Point #3: Roles are More Stable

As a consumer, I may be reading the news one minute, shopping for my family the next, and then making business travel plans.  Every organization I interact with has a different view of my journey and interests. While employees play different roles within an organization, they tend to have a job focus, be it marketing, compliance, product development or sales. Their role is more stable, and the content they engage with is driven partly by their job description.

For that reason, concepts like “engagement” and “relevance” take on a different meaning when you’re acting as an employee rather than as a consumer. While we always want to promote engagement, the meaning of engagement may vary dramatically depending on the person’s role within the organization. Someone in product development or training may produce and share a lot of content, while someone in HR or compliance may simply consume it. Does that mean that HR and compliance are not as “engaged?” Not at all. It simply means that their need for collaboration varies according to their role.

Charlene Li’s “Making the Business Case for Enterprise Social Networks” includes an excellent framework on page 13 to help you understand how social collaboration creates value for organizations. Here are some thought-starter questions based on her framework meant to inform your thinking about what to measure:

Encourage Sharing

  • Does participation in social collaboration correlate with employee satisfaction?
  • Does it promote best practices that lead to quality improvement?
  • Does it correlate to reductions in issue resolution time?
  • Does it correlate to an improvement in audit scores or other measures of risk reduction?

Capture Knowledge

  • Does collaboration shorten learning curves and/or training time?
  • Does it lead to reductions in onboarding time?
  • Does it translate to competitive advantage?
  • Does it reduce operational costs and better leverage employee talent?

Enable Action

  • Does it correlate to improvement in top-line revenue via customer retention?
  • Does it enable process improvements that lead to reduced costs?
  • Does it help speed products or services to market?

Empower People

  • Does participation correlate with employee loyalty?
  • Does more collaboration correlate with stronger employee performance and/or promotions?
  • Does it correlate with advances in innovation?

We are just scratching the surface of social media and enterprise social measurement, but I hope this spurs your thinking about the value of social collaboration to your organization. As always, I value your comments and contributions.

Posted in Uncategorized | 3 Comments

SAP on HANA: Real-TIme Enterprise is Looking More Real

SAP_Business_Suite_Powered_by_SAP_HANA_Events_2013_004_1Thursday’s news–that SAP announced the availability of SAP Business Suite powered by SAP HANA–could easily go unremarked by people outside the enterprise software world, but it’s an important milestone for those of us interested in big data, the real-time enterprise and, ultimately, predictive analytics. In a nutshell, the addition of HANA enables SAP’s business applications to capture and analyze transactional data in real time. (For a good summary of the announcement, read Rachel King’s post in ZDNet.)

What makes this announcement notable for those of us interested in big data, measurement and predictive analytics is the potential of an organization able to perform analysis on data and deliver relevant results at the moment of transaction. One particularly interesting example is BigPoint‘s Battlestar Galactica, a massively multiplayer game based on the immensely popular SyFy series of the same name.

Like many such games, Battlestar Galactica sells virtual assets as part of its monetization strategy. When a significant event occurs–i.e., you are shot–the system notes that and immediately tries to sell you virtual goods to prolong your gameplay (and, of course, deepen your emotional and financial investment in the game).  This is where we move into big data territory; in Battlestar Galactica, more than 15 million players are shot an aggregate of 5,000 times per second, so the game requires a propensity model to determine the appropriate offer based on the specifics of the event, the player’s gameplay skills, and other factors. To minimize gameplay disruption, all of this must occur–analysis to offer–within three seconds.

Another such example is the pre-paid phone card business.  What if companies could send a text message to the owners of pre-paid cards letting them know that their cards are about to expire, and offering the option of replenishing the card before it runs out? Pricing, promotions, decision support, supply chain: there are just a few examples of the countless applications for real-time (or, what some now call “right-time”) analytics.

SAP’s announcement also begins to reveal some of the massive changes in thinking that a so-called “real-time enterprise” will catalyze. For example:

  • Redefining the transaction. What is a transaction from a technical perspective? When you have the ability to do real-time transaction processing, it’s not just the outcome (buy a bigger spaceship) but the precipitating event (getting shot) that is important. This places much more emphasis (and yields more opportunity) throughout the entire customer journey, rather than the simple act of conversion.
  • The death of reporting. Rather than a culture based on reporting, this type of technology means that anyone in the organization could potentially query data, analyze it and make recommendations based upon it. Reporting then moves from an after-the-fact event to a continuous process, which could trigger some interesting cultural implications as accountability also becomes more “real time.”
  • The ability to predict. Alerting becomes more relevant. Algorithms can analyze massive amounts of similar situations based on historical data, run propensity models and provide weighted recommendations based on various scenarios.
  • Mobile becomes more strategic. Mobile devices would have the potential to query and act on unprecedented amounts of information, which requires more governance of, and potentially investment in, the mobile platform.
  • Design thinking becomes business critical. To ensure that customers get off on the right foot, SAP has included design thinking services for its initial customers to help them ideate and then select the business use cases that demonstrate the highest value and technical feasibility.  To make a real-time organization work, design thinking should become a core skill rather than a periodic event.
  • Developers design new apps based on real-time data. Said Amit Sinha, VP Database and Technology Marketing, “Our goal is to move the tool out of SAP’s hands and into the community.” The ability for developers to design apps for real-time event processing will open up new possibilities for business and consumer applications.

But before we get too carried away with rosy images of a real-time future, we must also look at the realities, many of which will follow my favorite law: the law of unintended consequences. While SAP executives stressed many times that SAP on HANA comes with minimal technology disruption, that doesn’t begin to account for the processes, people, policies, communication, decision making capability, culture and other factors that a truly real-time enterprise demands.  It’s not just different technology, but different thinking, different skills, potentially different cultural values that add up to a huge disruption in the way we value time and timeliness within an organization.  Of course, each company can control how quickly or slowly it moves toward this new model. Said Sinha, “It’s up to you to decide how much you want to adopt.”

That is certainly true today, when we are in the embryonic stage of real-time business. But, as social media harshly exposed organizations that underestimated the impact of an empowered public, real-time business will spotlight those companies and organizations whose processes are out of step with the availability and velocity of big data.

We will lose whatever patience we still have for experiences that neglect to incorporate the continuous streams of data that we create: restaurants who don’t “remember” that we’re regulars, retailers who consistently show us ads for items we’ve already purchased from them, hospitals who can’t let us go home for eight hours because their discharge process takes longer than our medical treatment. The fact that some organizations can behave in a more agile and customer-aware way will make those who can’t look sluggish and out of step by comparison.

Of course, this is a lot of “ifs.” If SAP is able to share successful use cases and prove market momentum early on, this will up the ante for organizations–end users and technology providers alike–to hasten the transition to a faster, more data-aware business infrastructure.  If customers adopt technologies that support real-time business and demand them from the the other vendors in their ecosystem, if the technology proves cost-efficient and reliable and transformative enough, well, you get the picture. But this announcement does throw down the gauntlet, for SAP as well as its competitors, to move past technology buzzwords and deliver real examples of real organizations creating and delivering real value.

To be sure, one thing we can rely on is that there will be disruption in both directions: for the laggards as well as the early adopters.

As always, I’d love your thoughts. Please share them in the comments, or elsewhere, and I’ll be sure to link to you.

Posted in Analytics, Big Data, Predictive Analytics, Real-Time Enterprise | Tagged , , , , , , , | Leave a comment

Big Data: It’s time to start thinking about DX

961774_58965270Back in September, I spoke on a panel at Dreamforce in which I used my Nike FuelBand to illustrate the difference between data and metrics, arguing that the difference is meaning: data is the raw material; metrics are one way you can use that material to create meaning. Here’s an example.

While counting calories was popularized in the early 20th century, the concept of a counting steps is relatively new. A couple of years ago, if you asked me how many steps I walked per day, I would have had no idea and, frankly, would have wondered why you cared. Then people started talking about the health benefits of walking 10,000 steps per day, and it became meaningful to a segment of the culture. While it’s not a “magic number,” and I am not confident that my FuelBand tracks my steps with complete accuracy, it’s not a bad barometer either. 5,000 – 6,000 mean I need to get moving; I really feel the difference when I’ve logged 12,000 or more. To some extent, steps walked per day has become a new health metric, at least for me.

Last year, Altimeter Group introduced the three themes we are focused on from a research point of view. The Sentient World is particularly intriguing, as it encompasses the implications of a world in which many devices–not just your phone, but your car, your carpet, your refrigerator, your fitness device or even prosthetic–collect and transmit data streams. And it’s not just devices, but the apps and platforms and APIs that collect and aggregate and analyze and share–that help to create the phenomenon we think of as “Big Data.”

There are many fascinating and worrisome implications to Big Data. On the consumer side, we worry first about privacy, then about being overwhelmed with more data than we can handle. We want accountability, relevance, control, utility, signal-to-noise ratio, ease of use. On the business side, we worry about liability, cost, processes, training, signal-to-noise ratiocost, analysis, integration, ownership and ROI. And cost.

All of us–whether in our business or personal lives–want what so many of my clients call actionability, or that nugget of insight in the data that tells us what to do next. Or, to put it another way, we are becoming increasingly overwhelmed by raw material, and increasingly in need of meaning.

The default way to solve this challenge is to give the user more stuff: more reports, more dashboards, more alerts, more badges, more notifications, more sizes and color choices.

But what I want, and what I hear from nearly everyone I speak to, is more meaning. More insight. More understanding of risks, opportunities, context. More guidance on what to do next.

The intentions are good, but the experience is broken.

Maybe we are at a point, as the Internet was several years ago, where we need to create a discipline around the experience and meaning of data.

Unlike information design or user experience design (of which I am a huge fan; don’t get me wrong), data experience design would encompass the experience of using many types of  data across channels, platforms and in real life. In that sense, it wouldn’t be a replay of UX, but a multi-dimensional discipline that articulates the rights of users, the criteria for “good” data–the methodologies, confidence levels, presentation, usability, privacy guardrails, governance, transparency, implications–everything we need to think about to make meaning from the gathering yottabytes and whateverbytes that continue to descend algorithmically upon us.

Whatever it looks like, whatever form it ultimately takes, one thing is clear: we need to jump off the hamster wheel for a moment and start to think systematically about the impact of big data and data science so we don’t end up creating a larger, more intricate and expensive replica of the past.

Posted in Altimeter, Big Data, Multichannel, Predictive Analytics, Quantified Self, Social Analytics, Uncategorized | 2 Comments

Binders Full of Women: When a Meme Hijacks Your Brand

During the second Presidential debate this week, it only took a few moments for a social media/community manager named Veronica De Souza to claim the URL bindersfullofwomen.tumblr.com. Not long after, YouTube exploded with parody videos. All of this is to be expected in the increasingly volatile, meme-hungry world we live in.

But at the same time, something started happening that affected more than the Presidential campaign: reviews of binders and trapper-keepers–lots of them–started spreading across retail sites, soon to be shared via Twitter, Facebook and other social tools.

Whatever your politics, the speed of data these days means that companies need to be on the alert not only for anticipated feedback about service and products, but for external factors–often impossible to predict–that can affect their brand, whether positively or negatively. These comments are a good example:

“I’m VERY disappointed in this binder! I had to order it one size larger than I am used to squeezing into – and it looks terrible with my pearls!”

“I’ve done everything I can think of to make this binder comfortable, but it is just too small, even for my delicate frame.”

You get the point. The challenge in this case is that many of these reviews are clever enough to include the product features in a subtle way. As a result, review tools need to include deep enough text analytics and/or human moderation to be able to distinguish legitimate reviews from those intended just for political and/or comic value. Then of course, the question is what to do about it.

I asked Pehr Luedtke, Vice President at Bazaarvoice, whether he had seen this issue arise among his client base, and, if so, what would or could be done. He said:

“We have not had any clients discuss a business impact with us regarding binder reviews. We do use human moderators that moderate content 24×7 to stop irrelevant and fraudulent content from going live on a customer’s site. We also use a fraud platform to detect fraudulent reviews to stop them before they reach our human moderators. At this point in time our fraud platform has not picked up on anything extraneous or irrelevant related to this. However, there could exist certain circumstances in which a customer uses clever wording to make the content relevant. But at this time we have not received any issues from clients regarding the implications of the binder comment on their business.”

Of course, some companies might see this moment as a brand opportunity. My colleague Rebecca Lieb, who covers digital media and advertising, had this to say:

“Binders are top of mind as a result of this speech in a way they probably have never been before and never will be again. So this is an oportunity for brands such as Avery to capitalize on this top-of-mind awareness. They can do this in a way that is thoughtful and based on buidling awareness and telling a story. They can even incorporate the fact that Halloween is coming up and people are already talking about dressing as binders. It’s a way for brands to be relevant in a tongue-in-cheek way.”

Of course, a topic as intrinsically polarizing as politics can be a double-edged sword, so Rebecca cautions that brands also must be aware of and plan for the unintended consequences of their keyword strategies. Witness the following; it’s not possible to tell whether it was intentional or not, or even whether it was based on contextual ad placement (given the likelihood that a company such as Office Max would likely purchase the keyword “binder”) or re-targeting (given that I did search for the word “binders” while preparing to write this post.)

Whatever you think of this or other memes, the “binders full of women” moment has a lot to teach us about the opportunities and perils inherent in digital and social media. Here are some recommendations on what you should take away from the “binders” brouhaha:

  • Make sure you have a process by which you determine what kinds of content you will allow on your site. A terms of use policy is a must-have.
  • If online reviews are important to your business, make sure your reviews solution is sensitive enough–and/or has human moderation–to block undesirable content from your sites, and/or to flag questionable content.
  • Use this experience as a catalyst to review your keyword strategy. If one of your keywords becomes a meme, make sure you control whether and how it appears in your portfolio.

The Internet is fertile ground for the law of unintended consequences, and you can never fully predict what odd confluence of events will affect your brand, intentionally or unintentionally, negatively or positively. Make sure your monitoring, content and ad targeting strategies take this into account, so whatever happens, at least you’re in the know, and in the driver’s seat.

As always, I welcome your comments and contributions.

Posted in Amazon, Crisis, Listening, Sentiment Analysis, Social Analytics, social commerce, Social media, Social Media Risk, VoC | 3 Comments

The New LinkedIn: It’s (Mostly) About You

Today, LinkedIn announced a revamp of its profile pages around three core principles: simplify (the experience), grow (in terms of network, value add) and everyday (provide everyday utility)

In April, when the company announced its iPad app, they described their thinking about how to optimize the platform for mobile in terms of usage statistics, which show that most people use LinkedIn in the morning (the “coffee” session) and at the end of the day (the “couch” session). This signaled a bit of a turning point to me that I hadn’t really heard from them before: the awareness that the company really needed to think about the platform in a more user-centric way, and turn it from a chore into more of a pleasure, frankly.

That’s why today’s news is quietly important: the profile page is simpler, more visual. New blog content by thought leaders–from Arianna Huffington to Marcus Samuelsson to Altimeter Group’s very own Charlene Li–is intended to draw us in and keep us there. The ability to tag ourselves with skills–and, better yet, have our contacts endorse us–creates a much richer discovery mechanism for potential contacts and employers to use when they are looking for people with particular skills; say Ruby-on-Rails or data science.

So, in a nutshell, I see two things that are slowly unfolding: content, and connectedness. Both are critical for LinkedIn to build virality and effectively use and monetize the vast amount of data at its disposal.

What it Means for Individuals

Have you ever had a recruiter call you and say something like, “I see you have three years experience as a blah-blah product marketing manager at YourCompany. I have a terrific opportunity to be a blah-blah product marketing manager at AnotherCompany.”

Yes? And did you respond something to the effect that it was nice of them to call and everything, but, all things being equal, why would you leave one perfectly good job for a lateral, nearly-identical job?  In a world characterized by emerging, fast-moving industries (think data science, for example), what you know and can do is as or more important than title and last position held. Especially if you are at all ambitious.

So it’s really becoming important to trick out your profile with your skills, especially because contacts can endorse you for them. This takes the cumbersome search aspects of LinkedIn and makes them a little more friendly. Imagine what happens when your profile contains a long list of skills, endorsed by your (high-quality) network? It becomes much easier to find you, and much easier to visualize your aptitude for that new, different position, based on the fact that you have a huge number of relevant and transferable skills endorsed by your community.  Here’s a preview (see full sample profile at the end of this post):

What it Means for Business

Of course, if you’re looking for someone with that elusive mix of skills, or trying to understand what the most commonly correlated skills are with, for example, big data, data science or something equally emergent, now you–and your recruiter–have the ability to better understand what aptitudes may also be relevant and transferable. Says Brad Mauney, Senior Product Manager at LinkedIn, “The endorsements have almost become a collaborative filtering for the things you know and the things you should know for a specific job.”

To me, this is the most intriguing aspect of the redesign; it turns LinkedIn from a lagging indicator (what you have done) into a leading indicator (what you could do). That has the potential to be an extremely powerful asset–for individuals as well as institutions.

The revamped profile doesn’t completely satisfy my thirst for effective data visualization yet–there are still plenty of cumbersome steps to go through to get the information you need–but it does demonstrate that LinkedIn is finally focusing on its most important asset: you.

Posted in LinkedIn, Predictive Analytics, Quantified Self, Social Graph, Uncategorized | 1 Comment

Social Media Crisis Prevention: Can You Defuse an F-Bomb?

It’s a nightmare scenario.

  1. You get a frantic text or call from a co-worker that someone tweeted a tasteless joke or profanity from your corporate Twitter account.
  2. The keywords in the post start trending among your brand mentions, soon to be followed by the hashtag #fail.
  3. Your customer service team receives outraged calls complaining about the foul language and lack of judgment of your employees.
  4. You scramble to delete and/or apologize for the offending tweet. The apology tweet must be vetted by at least three departments, delaying the apology and allowing the issue to escalate further.
  5. You are summoned to your boss’ office to explain what happened, what’s being done to fix it and what your plan is to prevent this kind of thing from happening again.

Sound familiar? It does if you are KitchenAid or StubHub, each of which experienced a similar social media crisis during the last few weeks.  But, even if you’ve been lucky enough to avoid this situation, you still need a plan.

The frst thing you need is a social risk management framework to guide you through the process of identifying, evaluating, mitigating and assessing emerging risks from social media. My colleague Alan Webber recently published “Guarding the Social Gates: The Imperative for Social Media Risk Management,” for that purpose. It covers the steps you need to take to understand your risks, and put the right policies, processes, resources and training in place to mitigate them. As Brad Hamilton (Judge Reinhold) says to Jeff Spicoli (Sean Penn) in Fast Times at Ridgemont High:

“Learn it. Know it. Live it.”

If you are in the healthcare or financial services industries, chances are you’ve been over this territory with a fine-tooth comb, given the stringent legal requirements related to patient, consumer and investor protection. But, if you’re in a less regulated industry, you still need to take stock of your social media governance policies and processes to prevent this–and other–scenarios from happening.

Clara Shih, CEO, Hearsay Social, agrees. “These [accidental posts] are certainly tactical challenges we need to figure out, but they signal that social media is growing up into something that’s real. Along with that comes a new set of responsibilities. If we expect social media to drive business impact, then we need to invest in a full set of governance policies, procedures and tools to make that so.”

Everybody else, this means you.

But what if you already have a social media risk management framework in place, but someone still makes a–frankly stupid–mistake and accidentally tweets something from the corporate account that they meant to tweet from their personal one?  Or if you have good processes in place to handle industry issues, but not cultural or political ones? Is there a last line of defense? Can technology be your safety net?

The answer is…sometimes.

Even if you are not in a regulated industry, it’s time to educate yourself about the social media management and compliance tools that include features and integrated business rules that can help prevent the unprintable. Typically, they include lexicons of industry-specific words or phrases (think “guaranteed return” for financial services) that require approval, modification or deletion to keep the brand in compliance with applicable laws.

In the case of StubHub, a profanity filter would have done the trick, assuming that the company had uploaded the most common swear words to their “flagged content” lexicon (and further assuming the tweet in question was posted by an employee and was not the result of  hacking).  In the case of KitchenAid, however, it’s a bit more complex, since the post did not contain any profanity but rather consisted of an offensive comment about the president and his grandmother.

[And let's be clear: the StubHub tweet was about the brand, by an employee, so would have violated most social media policies, irrespective of which account tweeted it. The KitchenAid tweet, while offensive, didn't enlist the brand except in the sense that it was tweeted from the corporate account.]

140 Characters of Prevention

Here are some quick tips on “last lines of defense” for social media risk. They can’t–and won’t–replace good policies, processes, education and judgment, but they can sometimes provide a safety net when all else fails.

  1. Posting restrictions. Look for management and compliance tools that enable you to set a requirement that people can only post to your brand account through their app, which would then run business rules that flag questionable content and/or prevent it from being posted in the first place. Make sure this is true both for the mobile and the web application, and change login credentials regularly to ensure that any employees or agency staff who have left the company no longer have access.  If you don’t have a tool that supports this function, Alan Webber counsels that you not allow employees to have personal and corporate accounts on the same mobile device (or app).
  2. Manual approval. While this is the most cumbersome option, and most difficult to scale, some organizations require approval of all posts before they go live on brand channels.
  3. Updated lexicons. Even if you already have a compliance tool with an industry-specific lexicon, consider creating lexicons of other types of language that you want to prevent or restrict on your corporate social sites. This could match existing acceptable use policies related to profanity, or discussion of politics, religion or race, if desired.
  4. Proactive monitoring. Look for tools that continually monitor your brand accounts and, depending on the content, automatically delete inappropriate posts or alert someone within the organization when they occur. Some of these tools also include infraction tracking, so you can monitor trends among departments, geographies and individual employees, as needed.

The truth is, you can’t always stop stupid. But for those times when judgment and impulse control fail, make sure you have the tools and processes in place to defuse that ticking F-bomb before it explodes all over your news feed.

For more on social media management and compliance tools, please see Alan Webber’s “Guarding the Social Gates: The Imperative for Social Media Risk Management,” page 20, as well as Jeremiah Owyang’s “A Strategy for Managing Social Media Proliferation.”  And, as always, I welcome your comments and differing points of view.

Posted in Social Analytics, Social media, Social Media Risk | Tagged | 2 Comments