eMetrics San Francisco 2011 — Recap by the Tweets

By on in , , , with 4 Comments

Note: There’s a lot of gushing I could do about how great it was to meet a lot of people in person whom I’d only known via Twitter prior, to see people I’ve met before, and to meet new people…but I’ll save some of that for a later post. This is the “content recap” post.

The last 3-4 conferences I’ve gone to, I’ve used Twitter in lieu of a notepad for my note-taking. What I realized after the first time I tried this was that it forced me to be succinct and to be selective as to what I noted. Now, for better or worse, my thumbs have gotten a bit more nimble at the same time that the input mechanisms on my mobile devices have improved. So, some might say I’m not as selective as I should be!

But, after eMetrics in D.C. last fall, I realized that there’s another benefit of tweet-based note-taking at conferences — it enables crowdsourcing the key takeaways. In theory, at least! Given that, I decided to organize this recap based on one thing: the top most retweeted tweets during the conference, as reported by a TweetReach tracker I set up. Scroll to the end of this post to download a CSV with the raw data, if you’re interested in crunching it yourself.

With that, here are my six summary takeaways:

The Data Isn’t Actionable without Storytelling

Hands, down, the most retweeted tweet (14 times) from the conference was this from Wendy Ertter:

Several presenters touched on the fact that one of the key challenges in our industry is communicating what the data means. As analysts, it can be easy to get absorbed in the data to the point that we intuitively can interpret our analyses. All too often, though, we forget that the business users we’re supporting are neither “wired for data” nor have they been as immersed in it as we have. So, rather than getting stamp-your-foot irritated that your brilliant insights have not led to action, take a look at how those insights are being communicated.

Now, not discussed was the fact that “tell a story with the data” can easily come across as “torture the data until it tells the story you want it to.” It’s a fine line, really, that means transparency has to come along with the storytelling. Storytelling must be merely a means of “effectively communicating the truth” — conveying what the data really is saying, but in a digestible manner.

Social Media, Social Media, Social Media

Ken Burbary’s tweet during Guy Kawasaki’s closing keynote (which garnered quite a bit of ire from the attendees, but that’s potential fodder for a future post) was retweeted 11 times:

Social media was a hot topic at the conference, with the sessions devoted to it concluding: “It’s tough to analyze.” In general, there was consensus that Performance Measurement 101 still applies — if you want to have any hope of measuring social media, you darn tootin’ better have clear objectives for your investment in the channel. Now, because social media isn’t the same as longer standing channels, there are different measures to work with.

One of the more intriguing sessions I attended was a panel, moderated by Michele Hinojosa, that featured Gary Angel of Semphonic and Michael Healy. The subject was sentiment analysis. Specifically, sentiment analysis of short-form text messages — Twitter and the like. Both by illustrating examples and talking through some of the advanced machine learning algorithms that have been applied to the challenge, they made a pretty strong case that trying to discretely quantify sentiment in a Twitter world is a fool’s errand.

Gary also made a distinction between “monitoring” and “measurement” and, later in the discussion, postulated that social media may be one case where you actually need to do analysis first and then set up your measurement. This makes sense, even in light of my “Performance Measurement 101″ comment above. It does make sense to sift around in the conversation that is going on around a topic or a brand a bit to get a human and qualitative sense of the lay of the land before determining exactly what to measure and how.

[Update: I just realized that Gary wrote up a pretty detailed post about his key points in the session over on his blog last week — it’s worth a read.]

Attitude Is As Important As Behavior

This tweet from @SocialMedia2Day during Larry Freed’s opening day keynote was retweeted 10 times:

Foresee Results was the Diamond Sponsor for eMetrics, and the company continues to push the web analytics industry to recognize attitudinal data as being every bit as important as behavioral data. Interestingly, VOC vendors overall had a much more prominent presence than web analytics vendors (only Google Analytics and Yahoo! Web Analytics were exhibitors at the event — Webtrends, Adobe/Omniture, and Coremetrics were nowhere to be seen in the exhibit hall).

I have to credit Chris Dooley from Foresee Results for initially introducing me to (read: pestering me about) the rightful place of attitudinal data as a companion to behavioral data. He was right when he started preaching it, and he’s still right today. Another VOC vendor noted during his presentation that, when his company surveyed the top 500 retail sites and the top 500 overall trafficked sites, they found that only 15% were running on-site surveys. That is both surprising and alarming! OpinionLab also impressed a number of people with their presentations in the exhibit hall theater, and iPerceptions provided a bit more detail about their coming 4Q Premium product (which, seeing as how they announced it was coming back in October, is somewhat underwhelming given the price tag).

In short, lots of reinforcement that the voice of the customer matters and shouldn’t be ignored!

comScore’s Silver Bullet (A Bit Tarnished, IMHO)

Since I said I’d go with the most retweets, I have to include this one from John Lovett, which was retweeted 10 times:

The key here is that comScore announced all the problems they were solving. The main differentiator, as best as I can tell, is that comScore is combining web analytics capabilities with its rich demographic/audience-based data. That might be slick, although it seems that they’re overpromising a bit when it comes to the flexibility of the tool and the completeness of the demographic data. I trust John…a lot…so maybe I’m being unduly and prematurely cynical. We’ll see.

Consumers Are Cross-Channel — So Should Be Your Analysis

At the risk of inflating John’s ego (which I’m not all that worried about, but if, ages hence, he’s turned into a pompous ass, I’ll dig up this post and claim credit for starting a perfectly pleasant guy down that path!), the next tweet and the last one are all Lovett-related. Lovett. Love it! :-) This next one was Eric Peterson quoting John and was retweeted 10 times:

Data integration and cross-channel analytics were covered by a number of presenters. With the exception of one vendor (who shall remain nameless…but who announced a name change to his company at the conference), the overwhelming agreement was that cross-channel integration is hard, tedious, expensive…and necessary. That one vendor had a video that showed it as being simply a technology issue (and they had the technology!). I’ve dabbled in the customer data integration (CDI) world enough to know that doing this integration at the individual person level is a bear.

But, because customers are living in multiple channels — offline, digital, mobile, social — and are switching freely between them, it’s dangerous to narrow in on a single channel and draw too many conclusions. This challenge isn’t going to go away any time soon.

Several times, both in sessions and in hallway discussions, it came up that both “WAA” and “eMetrics” have quickly become misnomers. Most of the attendees at the conference have responsibilities well beyond simply “web site analytics,” and simply “digital metrics.” I put a plug in that we could start considering “eMetrics” to be “everywhereMetrics,” which is a shameless ripoff of Resource Interactive‘s stance that “eCommerce” has become “everywhereCommerce.”

Fun times to come!

Consumer Privacy — the Regulations, the Law, the Ethics of It

Covered briefly in several sessions, touched on in the WAA Member Meeting, and then covered in depth in a panel was the challenges our industry is facing with regards to consumer concerns about privacy:

John has been the face of a multi-person effort to craft a code of ethics that individuals can sign that lays out how we will treat customer data. What became evident at eMetrics is that there simply is no easy answer to “consumer privacy.” And, the fact that the FTC covers the U.S. and has taken differing stances from the EU, and the EU will get to “one policy…implemented and enforced by country,” just makes my head hurt.

The good news, it seems, is that there seems to be an emerging philosophical consensus as to what is “good/okay” and what is “bad” when it comes to user tracking. The kicker is that it’s really, really hard to write that down in an unambiguous, loophole-free way.

If anything, I took away a sense of empowerment when it comes to really living the Code of Ethics and speaking up if/when I see an initiative starting to get into a gray area — it’s not just a “do the right thing because it’s ethical” case at this point. It’s a “do the right thing…or it might come out that you didn’t, and your brand can get burned severely.”

The Tweets Themselves…

As promised at the beginning of this post, if you want to download the data file with all of the tweets from 14-Mar-2011 to 16-Mar-2011 (Eastern time) that came out of my Tweetreach tracker, you can do so here. If you do anything interesting with them, please leave a comment here as to what that was.


  1. Great content recap Tim.

    Let me get out of the way that it was a pleasure meeting you last week.

    The best part of these types of posts is I can say, “Yeah, what he said.”

    I had similar takeaways most notably is how to use storytelling to make data actionable. I feel that the importance of this idea cannot be overstated. This does not have to be taken as far as some of the brilliant examples in some of the presentations. Simply empowering hypotheses using data rather than spitting out spreadsheets is a takeaway worth the price of admission.

  2. Pingback 2011 Spring Break Part 1: San Francisco Sans Kids | Second Tree Blog

  3. Pingback One Digital Analyst’s Guide to Using Twitter | Gilligan on Data by Tim Wilson

  4. Pingback » Monish Datta Gives #cbuswaw w/ Eric Peterson & ForeSee Thumbs-Up | Tim Wilson's Blog at Web Analytics Demystified

Leave your Comment

Notify me of followup comments via e-mail. You can also subscribe without commenting.

« »