Final Thoughts on the Berlin X Change
When I did my wrap-up on X Change Berlin, I went through my notes and, in one of the concluding paragraphs, I listed some of the topics I intended to discuss in more depth. After having covered Persona-Based Segmentation and Reporting Fatigue, however, I realized that at my current pace I would be finishing the series on Berlin learnings well after the next X Change in LA! So I decided to collapse the schedule somewhat and wrap things up with a final post that covered some of the most important take-aways I had from the Conference.
Let me start with something that came up in the Keynote and that really got me thinking about building digital teams. David McBride mentioned that his effort in Big-data analytics got a tremendous boost when he brought in someone with a strong traditional background in customer analytics and database marketing. Looking at our panel of successful big-data enterprises, it struck me that they had all somehow captured that blend of expertise and I heard much the same thing in conversation with David Williams of ASOS (who led one of our Big Data Huddles). If you’re current digital measurement team lacks that set of traditional customer analytics skills, it’s going to severely limit your ability to succeed in the analytics warehousing world. It’s partly a matter of tools & technology, but really it’s much more about method and approach. After all, David’s team pursued a cloud-based digital analytics solution that, I assume, was pretty much outside everyone’s immediate technology expertise.
That’s also a key role I see for Semphonic. For almost two years now, our basic capabilities deck has features a couple slides that talk to our role in driving exactly that convergence in digital measurement:
It’s a long and often painful process transitioning from offline analytics to Web analytics to digital customer analytics. I see organizations try to drive this transition in almost every conceivable fashion: from kicking digital data over to customer analytics teams, to creating big data platforms for digital teams, to leavening digital data teams with customer analytics expertise. Based on what I’ve seen and based on what I heard in Berlin, I’m more and more convinced that this last approach is probably best.
Tag Management has been a very popular topic at the last few X Change sessions and Berlin was no exception though the level of hands-on experience there was much more limited than I think we’d get in a corresponding session in the U.S. One of the surprising aspects of the discussion was the interest in Build-vs-Buy when it comes to Tag Management. There are some reasons why a company might consider a build decision in Tag Management: the existing vendors are small and TMS represents a level of technology risk, it isn’t actually that complex a technology and it’s fairly straightforward to implement some form of Tag Management, and the addition of another layer of cost is unpalatable to many enterprises. However, all the many reasons to lean toward a Buy decision are also in place here. This is a rapidly evolving technology space and keeping up with the analytics vendors is going to be an ongoing commitment. Internal development is always harder, more time-consuming and more expensive than you think. The actual cost of Tag Management can be quite low – and there are enough vendors to choose from that you should be able to drive a palatable bargain. In addition, the TMS vendors are increasingly building GUI-based functionality into their systems that is going to be difficult or impossible for your internal teams to match.
Joe Stanhope of Forrester Research gave a thorough summation of these issues and left me, at least, in little doubt that “Buy” is the right direction for almost any enterprise in this space. Oddly, the one place I think a “build” oriented approach might make sense is around a CMS-based solution such as CQ5 – so it really isn’t a “build” approach at all. As I mentioned in a previous post, Adobe’s best Tag Management isn’t being done in their own Tag Management Solution – it’s in their CMS. And I’ve long been a believer that the CMS is the right place (ultimately if not in today’s market) for Tag Management.
Lastly, I wanted to talk a bit about a couple of really interesting directions that we discussed in Alex Emberey’s Huddle on Social Media measurement.
One of the great challenges in Social Media measurement is finding ways to assess the impact of social efforts on sales efforts. You can track direct linkages, but this probably reflects only a fairly small percentage of your overall impact. You can also track volume of social media linkages by aggregating all social referring sources together. I thought Alex had a great slant on this latter approach by measuring the quality of aggregated social media traffic vs. social media campaign efforts. If you’ve setup your Marketing Channels or Campaign reports correctly, that’s fairly easy to do in a Web analytics tool. By careful comparison of social media effort and traffic quality for specific periods, you have a pretty good tool for assessing not only how much traffic you drove, but whether or not that traffic was actually business useful. This approach doesn’t provide any deep granularity – you can’t assess individual campaigns – so it’s almost like a sub-class of media-mix modeling. What I think makes it valuable is that you’re actually getting a strong read on the degree to which social efforts are tying to real business.
Another extremely challenging aspect of Social Media that is ultimately critical to its effective use is finding ways to take Social Media from a broadcast channel to a one-to-one channel in a scalable fashion. Social channels are inherently personal and are only going to be really effective when used in that fashion. But achieving any level of “personal” communication at scale is incredibly difficult and almost always cost-prohibitive. Because of this, most enterprise’s have chosen to treat Twitter and Facebook and LinkedIn as essentially ‘blast” mediums. One of the most interesting discussions in the Huddle revolved around the creation of semi-automated response templates. These would categorize and score posts with certain highly respondable subject categories for which templatized responses were pre-prepared.
This type of system places some pretty serious demands on the ability of your systems to categorize and score conversations – part of the reason, I suppose, that I find it interesting. But finding ways to create scalable but personalized Social Media is at the heart of the Social channel challenge.
I hope these posts have been useful in their own light AND also that they’ve provided a sense of how rich X Change is as an experience. Despite many years in the industry and lots of hard-won experience, I never walk away from X Change without a head-full of new ideas and perspectives. If you joined us in Berlin – Thanks! – and I hope to see even more great European practitioners there next year. If you’re here in the U.S., this is a great time to book X Change Los Angeles in September. If you’re like me, you won’t be disappointed!
