I believe that unstructured text represents the biggest problem and opportunity in Big Data, yet both are poorly understood by many consumers of text analytics.

In my presentation I tried to shed some lights on the challenges facing text analytics, with most of the examples from sentiment analysis (as this is the easiest application to understand).

For those who were unable to attend, here is the presentation, sans my dynamic presence, of course

App developers who helped make Twitter more useful and accessible to got their first kick in the groin Twitter bought Tweetie iPhone app in April 2010. At that point, any company in the mobile Twitter client business was toast. About a year later, Twitter bought Tweetdeck and told developers to not “build client apps that mimic or reproduce the mainstream Twitter consumer client experience.” Now, desktop/webtop apps were toast too.

This was the beginning of the end of a cottage industry of innovative 3rd party apps that had helped Twitter become accessible, popular and more complete. Since then, Twitter has constantly changed the definition of “the mainstream Twitter consumer client experience” as it picks up successful ideas from 3rd parties and then takes measures to squeeze those apps out of the market.

The latest example of this is Twitter’s plans to create richer, magazine-like experiences via expanded tweets, inspired by apps like Flipboard, Instagram and YouTube. This richness can only be experienced at twitter.com and cannot be replicated by companies like Flipboard, who will only be allowed to peddle 140 characters or less of text.

3rd party developers are invited to the party again, this time to help create content for these ‘richer’ units on twitter.com. (When I say developers, I mean content publishers. App developers are running for the hills).

It seems that Twitter is taking a cue from Facebook, which successfully convinced brands and publishers that it was better (or at least critical) to set up shop inside of Facebook and treat your owned media properties as of secondary importance. I have no doubt that media companies would jump onboard – and I can’t blame them. It makes business sense for both publishers and Twitter.

However, for app developers there are lessons to be learned.

I become cognizant of the dangers of building on top of Twitter after my business partner and I built the “Realtime Pro Football” iPhone app in 2009 and Twitter shortly thereafter started going after apps that mimicked the Twitter user experience. While Twitter had yet to object (and probably wouldn’t have), I decided it was too risky to continue down this path.

Facebook felt like a safer place. When I launched “Open Audience Manager” in September of 2010 I felt more confident that I wasn’t building my house on sand. This product was an enterprise Analytics and Engagement platform for Facebook’s Open Graph. It helped online retailers make every product in their catalog an Open Graph object, and as such offered Facebook insights data at the product level, not just for a brand’s Facebook Page. Additionally, it made it possible to use Facebook’s communications channels to start a dialogue with Friends of each product – about that specific product, driving them through the purchase funnel. The product generated unprecedented ROI for the F100 clients that used it. Until now. Facebook just announced that they are removing the ability to publish messages to Friends of Open Graph web objects. They basically removed a key part of the value proposition of the Open Graph: Open Graph Objects can live anywhere, sharing a common set of bonds and communications channels with its Friends.

The problem with both Facebook and Twitter is that they haven’t completely figured out the role of 3rd party developers. They may say they embrace the development community, but they don’t really respect the community. Companies like Salesforce and Google have figured this out, mostly because they are committed to helping 3rd party developers make money as a way for themselves to make money.

That said, while 3rd party developers may be disappointed, this is not really a criticism of Facebook or Twitter as businesses. They need to maximize their revenue. A long as developers understand that Facebook and Twitter aren’t true development platforms, there is nothing wrong with building products and services that leverage Facebook and Twitter. I urge 3rd party developers to do so.

If you want to be part of the social media revolution, you can’t do it from the sidelines. You have to make your bets and take calculated risks that something that you rely on may be here today but gone tomorrow.

Good luck.

Google+ seems to be all the rage these days, with wildly divergent views among social media practitioners and users with respect to the viability of the social network. Personally, I’m still on the fence with regards to the current value of Google+ to brands and marketers, a view I shared with my fellow presenters and listeners via Bulldog Reporter’s PR University webcast yesterday.

For those of you who were unable to listen in, and for what it’s worth, here’s a summary of my perspective on Google+ as a marketing channel.

Friends are optional

Users don’t need to bring their friends to enjoy Google+. On Google+, users can seek out topics and then discover interesting conversations among strangers with similar interests. In that respect, it is more like Twitter than Facebook: Facebook has no value without your friends. Therefore, Google+ should be a good place to find and engage new audiences who may not know about you today.

Users will be drawn to Google+ through content experiences

The gateway to Google+ will be Google search, Google Reader, Google “Flipboard” (Google is allegedly developing a Flipboard-killer), Picasa, YouTube, Google Music and other Google services that drive content discovery and consumption. While these services are poorly integrated today, they will likely improve over time, and will become important traffic sources into Google+.

Google+ will become a hub for public, topic-based communities

Ning tried to create smaller, niche social networks, and largely failed, but Google may actually succeed. The combination of Circles and Search makes it easy to discover and engage a conversation around a topic. That is not easily achieved on Facebook because most conversations are private, and it is difficult on Twitter because there are no discussion threads to follow. For example, try these Google+ searches to see what’s on people’s minds within some hot topics, as well as fringe culture;

• “HPV” on Google+
• “Lego” on Google+
• “diy analog synthesizers” on Google+

Google+ won’t kill Facebook

People who use Google+ are not looking for an alternative to Facebook. They want a better experience on Google properties. Create more content that engages and facilitate conversations around your content.

Forget about brand pages (for now)

Have your employees engage and listen to figure out what works for your brand. Rules of engagement on Google+ are mostly similar to other social venues, so there is no need to revise your social media policies.

Google+ is important for SEO

It’s fair to expect that conversations and engagement around branded content and branded terms on Google+ will yield SEO benefits, just like signals from other social networks are already influencing search results.

To summarize the recent announcements, two major changes occurred on Facebook.

First, Facebook announced the Timeline. This is a chronological, user-curated view of every activity the user has taken or announced via Facebook that he or she deems meaningful. It’s a personal scrapbook annotated with photos, status updates, music, location check-ins, etc. This may become the most valuable real estate for users and for brands. Marketers may create applications that interact with the Timeline.

The second big change is the expansion of a piece of an underlying technology called the Open Graph. Previously the Open Graph allowed brands to include owned web properties into the Facebook ecosystem via the Like button. Now brands may go beyond the Like and embed custom verbs and actions into their applications — on Facebook or on their own sites.

This may result in more brand-relevant and viral newsfeed updates. Instead of simply Like-ing content on a brand site, consumer can send the following types of messages into their streams, from branded properties: “Joan is redecorating with the bold look of Kohler” or “Ron is grilling an Applegate Organic Hot Dog with Laurie.” Users can choose to include these branded stories or experiences into their Timeline, thereby elevating them to updates likely to be seen by friends.

Three Implications for Social Activation

First, Facebook wants brands and their customers to engage without forcing the experience to be optimized around Likes or Fans. Instead, marketers should use verbs that describe actions that make sense with their brand, and make it easier for these social experiences to live outside of Facebook.com.

Second, Facebook applications, which can be Facebook-enabled experiences on Facebook or branded websites, will become an essential part of the marketer’s toolkit. Facebook is a content distribution network that happens to come with a social graph. Therefore, the most viral experiences will be content experiences. This further increases the requirement that brands act more like publishers, and seek to create content that is engaging and shareable.

Finally, making it into the Timeline via Open Graph-enabled experiences will become key to marketers looking to create long-lasting relationships with their audiences.

Three Implications for Social Measurement

1. Benchmarking becomes more complex

As brands increasingly spread their Facebook investments beyond Facebook.com with unique experiences and action metrics, benchmarking one brand against another will be more challenging. Tracking competitors’ Facebook activities beyond the Page will be required to assess the merits of their tactics, and perhaps decode their strategy.

2. Measure what matters to you

The most valuable metrics for any brand to track are metrics that align with your strategy. With the freedom to create unique experiences and your own social “verbs”, you should focus on your specific Key Performance Indicators (which may be other than “Likes”) and track whether you are improving effectiveness around those. For example, if you are an online media company, you may want people to “read”, “watch” and “listen” more. If you are selling products you want people to share that they are “trying”, “buying” or ”drooling over” your products.

3. Measure. Iterate. Repeat.

Something that isn’t new, but worth repeating: Measurement is of no value unless it helps you improve your marketing. Metrics need to be at the core of all your marketing programs, and all your marketing programs and applications must be designed to be adjustable during their lifetime.

Social Beyond Facebook

Marketers must come to think of social as a behavior — rather than a channel — and look for ways to engage audiences in a brand-relevant way, regardless of which network enables the behavior. With the Open Graph, Facebook wants to be an enabler that doesn’t get between a brand and their customers. Marketers no longer need to use Facebook vernacular and metrics such as Likes or Fans in order to create viral and measureable experiences for Facebook members.

But Facebook isn’t the only game in town, especially when you start treating social as a behavior.

How will Facebook’s changes affect your marketing?

If this sounds like Twitter, it’s because it’s very much like Twitter. I suspect Google+ will follow suit soon with a similar concept.

I think this is a great feature, and would be even greater if it was available to Pages.

However Facebook will probably not make this available to Pages for a simple reason: Facebook doesn’t want to give marketers a short-cut into their Fan’s Newsfeeds. That’s what ads are for.

Lets see if Google is more business friendly.

Sometimes it takes a lot less effort to validate assumptions about your product-market fit than one might think. You can come by customer feedback earlier and with less effort using Lean Startup methodologies.

That’s what I learned last weekend, as my career came full circle last weekend during an intensive “Customer Development” workshop lead by a new entrepreneur-education focused startup called Lean Startup Machine. When I say full circle, I mean back to focusing on consumer insights.

Nearly 14 years ago I started working at Ogilvy as an Account Planner. Account Planning (in the US at least) was then a novel, but increasingly important part of an ad agency’s structure and workflow. In short, an Account Planner (also called Strategic Planner or Strategist in some agencies) represents the voice/interests/perspectives of the consumer and is responsible for bringing the target audience to life during the creative development process. It was in many ways a recipe that helped agencies live up to Harry McCann (of McCann Erickson) 1912 mantra that good advertising is simply “Truth Well Told”.

Over the last decade I’ve found myself in early start-up environments where marketing (positioning, messaging, creative, media) and the product development process has been one and the same. It’s all about finding a product-market fit and communicating that fit to the target audience. From the get-go that requires a deep understanding the customer.

I’ve typically enjoyed the start-up product development process more than just marketing. In my experience, marketing (and advertising) sometimes comes down to making a product nobody needs look desirable to whoever can pay for it. Wouldn’t it be better to build a product that people need in the first place? (This sentiment is echoed in Seth Godin’s book, Purple Cow, where he says that marketing remarkable products is ‘easy’ as they already have the marketing built into them.)

The “Lean Startup Machine” workshop was a great way to get re-focused on the customer when you’re in a startup environment. As an ex-Account Planner and with plenty of scars from startup adventures, the premise of this workshop was close to my heart. However, what I did learn was that sometimes it takes a lot less to validate assumptions about your target audience (i.e. how much they actually would need or appreciate your product/service) that one might think. I realized that on occasion I’ve spent resources on building technologies (proof of concepts, minimally viable products) unnecessarily, because I could have come by customer feedback earlier with less effort.

So whether you’re an Account Planner or a startup product manager, you already know you should be listening to your customers and prospects. If you’re curious as to how you might gain customer insights sooner and cheaper than you may think, check out the next class of the Lean Startup Machine or this retrospective post from one of the workshop’s mentors, Kelly Boyd.

Google+ Front-end innovations

• We’ll see new (or new uses of existing) music, video and photo services, both for consumers and content producers, powered by Google’s HTML5 based Wave technology for real-time sharing and communications. Expect Picasa and Google Music to be tightly integrated with Google+.
• There will be lots of social games, powered by Google Wave technology.
• Android and Google+ will become tightly integrated, and along with that, many new location-based services (discovery and advertising/commerce)
• Google+ Circles will become two-way: Right now you can place your friends into circles. This allows you to view aggregated streams by Circle, as well as you can publish to selected circles. I predict that Google will allow users (and businesses) to have multiple circles representing themselves, like sub-accounts. This will make it easier for a business with multiple locations to have a main page and a sub-circle representing each branch, or one circle for promotions, one for brand information, etc. Individuals can set up the same type of Circle segmentation, e.g. you’ll be able to subscribe to a friend’s “recipe tips” but you don’t have to get her Farmville, or Twilight updates.

Google+ Back-end innovations

• AdWords will become a behavioral ad network as social analytics become part of Google Analytics and the AdWords engine. This will in turn lead to further adoption of the +1 button.
• There will be ads on Google+, but more importantly, actions on Google+ and across the web via +1 buttons will create new hyper-targeted, behavioral advertising opportunities across all Google properties as well as the AdSense network. As a result Google Analytics become one of the most powerful/useful analytics packages out there (regardless of price).
• HTML5 grows up and kills of Flash

What do you think?

Sidenote: What is the impact of Google+ on other social media platforms?

Here’s my projection: Google+ won’t kill Facebook, but will challenge Twitter and micro-blog services like Tumbler and Posterous.

As a result Apple (the only major play without a real social offering) will buy Twitter, Hulu and Spotify (before Google or Microsoft does) which as a result will kill off MySpace once and forever.

While there’s a case to be made for the efficiency of short emails, there isn’t always one to be made for effectiveness, which I hope the following two examples will illustrate.

TLA galore (sorry, Three Letter Acronyms)

While email is not bound by the length and formatting limitations of mobile text, many correspondents still exhibit a spartan use of words with a gracious use of acronyms and emoticons. For example, I had an email exchange with a guy who simply replied to a meeting request with four words: “somewhere around UES EOD”. I figured out that EOD was End Of Day (I use that acronym myself) but it took me a while to figure out that UES was not the name or acronym of a company or a building, but simply “Upper East Side” (of Manhattan). Effective or efficient it was not.

Long live ADD (I’m not spelling that one out)

Secondly, I’ve found that an email is no longer great at addressing multiple issues. Many email correspondents seem to have adopted a threshold of one, in terms of how many streams of thought that can be expressed in any one correspondence. I find that if asking recipients to address more than one issue, question or deliverable, the reply tends to be a partial answer.  I’m learning slowly to ask questions as I would if I was texting: Only ask for one thing at a time, as multiple exchanges seem inevitable. So much for the efficiency of short messages…

Why email is becoming like texting

I think there are a few reasons why we’re seeing this convergence of protocol when it comes to texting and emailing. So, without making the fundamental attribution error (of blaming people rather than looking at the context), I think most people write short emails because they are composing it on their mobile phone.

That said, I know several people who have adopted the “Sent from my Mobile” signoff for their non-mobile email as well, just so that they can save a few keystrokes without coming across as rude. (We blame it on Kevin Rose!).

The difference between short and curt

While I always initiate email correspondence with a stranger with a ‘proper’ business etiquette, I pushed the boundaries on how soon to should skip the formalities the other day. This created an adverse reaction from the recipient of my emails and I understand why. Hopefully this story can be some food for thought.

Here’s what transpired: After my ‘proper’ initial email, the recipient replied with equal poise that she would like to talk, and suggested a time to speak and asked for a number to reach me on.

Here’s where things went awry. I was out of the office when I got this email on my cell phone. I wanted to get back to her quickly to secure the time slot with a quick to-the-point message (as I’ve learned that people’s calendars change by the hour).

My reply read:

4:30pm should work. You may call me on the number in my signature.

Sent from my iPhone

I hoped that hiding behind the “Sent from my iPhone” signoff would mitigate the risk of being perceived as rude, but rather as responsive.

It didn’t.

Fortunately, once we got on the phone, the incident turned out to be a great icebreaker and we got to learn more about each other and how we feel about business and business communications (including several of the points I’ve made in this post) as a result of it.

However, think twice before you decide it’s OK send short emails like this yourself. It may be the last email exchange you’ll have with that person.

The future of F-Commerce is off Facebook

Facebook’s Like button for websites is perhaps the most powerful long tail F-commerce enabler. It’s free and takes little technical effort to integrate across your entire product catalog. Whenever a Facebook member comes across a product with a Like button, a simple click makes them a “Fan” of that product, generating a Like Story in the News Feed of their friends. This is valuable as this free viral marketing channel not only creates band impressions, but also traffic back to the product page.

Spread the love

Depending on the amount of inbound traffic a product page receives, each product may only earn Likes in the single or double-digit range (although there are plenty examples of product pages with thousands of Likes). What’s important is the total number of Likes across the product catalog. These long tail Likes are very important because, other benefits aside, they each generate a high-visibility Like Story in the News Feed.

I would argue that for online retailers, Like Stories generated when Facebook members Like product (web) page are more valuable than Like Stories generated from Facebook pages. Consider these two Like Stories that appear in my friend’s News Feeds after I Like Dell’s Facebook Page vs. when I go to their website and like the XPS 15z Laptop.

If you were Dell, which message would you rather have me spread to my friends?

Joy in repetition

The other big benefit of Like Stories generated from the retailer’s website is that there is no limit to how many stories a visitor can trigger. Whenever I Like a product from an online retailer, a new Like Story is spread to my Facebook friends. Conversely, when I Like a Facebook page, that’s the end of my advocacy; it’s a single blip on the radar.

According to a February blog post by Blind Five Year Old, 28 % of the US top 100 online retailers have added the Like button to their product pages, and given the Like buttons’ rapid growth (10,000 new websites every day), I suspect the number to be higher now.

However, very few retailers have placed the Like button in prominent places on their product pages and are probably missing out on an incredible amount of free, viral exposure because of it.  (Good example of placement from Sephora, and here’s a no-no example from Dell where the Like button is buried below the fold.)

Once online retailers discover the strategic value of the Like button, they will re-design their pages to make the button as prominent as their own “Buy Now” button. At this point I predict online retailers will find that social traffic to their individual product pages drives sales more effectively than the traffic to their Facebook page. The implications for their media strategy should be significant.

That’s my 2 cents.

For my perspective on the Like Button’s cousin, the Send button, read Think Twice Before Implementing Facebook’s New  Send Button.

The short answer is yes, but Facebook’s Edge Rank algorithm is a “black box” and  in constant flux, which makes it hard to come up with a definitive list of tips and tricks, as I’ll illustrate toward the end of this post. However, there are a few best practices – which can be best understood by understanding the purpose of the Edge Rank algorithm. Before we get to that, let’s look at NFO and SEO in context.

Both NFO and SEO are about optimizing content in a way that maximizes its exposure and reach. However, Google’s Page Rank algorithm operates in a fundamentally different way from Edge Rank. For one, the Facebook News Feed is a push channel while Google search is a pull channel. Furthermore, Google doesn’t care who you are (although they are working on personalizing search), while for Facebook, your identity drives everything.

The purpose of Edge Rank

The primary purpose of Edge Rank is to protect Facebook members from information overload. Once you, as a Facebook member, reach hundreds of friends, Like dozens of Facebook Pages and hundreds of Open Graph pages, you may have several hundred stories queued up and ready to hit your News Feed on any given day. Edge Rank works like a filter that decides which handful of stories to “let through” based on how relevant and interesting the algorithm “thinks” you’ll find each story.

Edge Rank is calculated the following way: Weight x Affinity x Decay (time)

Weight = Weight for the edge type (create, comment, like, tag, etc). For example, publishing a post may be more important a like, sharing photo might be more important than a comment.

Affinity = For page posts (status updates), this is a score that quantifies the quality of the relationship between publisher and recipient of a story, e.g. how often members interact with content generated by the edge creator (publisher). I would assume that the number of common friends between the two parties plays into this as well, especially if those friends also have a high affinity score with the publisher.

Decay = How long ago the edge was created. E.g. Who cares about yesterday’s news?

Given that every recipient is different, it may seem impossible to come up with a way to optimize page posts, but it isn’t. There are a few rules of thumb that will help increase each post’s exposure. Just Google “Facebook News Feed Optimization” to educate yourself on the latest insights.

Is popularity a bad thing?

Like I said, Facebook’s algorithm is in constant flux, and your Google search may not uncover this interesting piece of information that showed up in my email inbox this morning. According to Page Analytics company, Pagelever, your Page’s Fan count affects your ability to break through to your Fans’ News Feed. According to the company, pages with:

• 10,000 fans reach about 30-40% of their fans per status update
• 100,000 fans reach about 20-30% of their fans
• 1,000,000 or more fans reach about 10% of their fans

It may seem, then, that the more popular your page is, the harder it will be to get your content in front of a large share of your fans. Sounds counter-intuitive, or maybe not: I suspect the underlying cause is that the most popular Pages in terms of Fans may also score low on affinity with their Fans.

Has anybody seen data to support or refute this?