I have been looking at the OpenAmplify semantic platform and examining how it may fit into Analytics implementations. How OpenAmplify works is illustrated in the example below
OpenAmplify looks interesting and sexy, particularly if it can be retrofitted to social listening platforms such as Radian6; other integrations exist, one of these is with Market Research company Millward Brown, has been using OpenAmplify since last year to help cut down on the manual coding of document verbatim (I'm assuming). Here's an example of using OpenAmplify that was produced by the company:
Millward Brown is happy with the results they are getting with OpenAmplify based on a Semantic Web post from late last year. I think the quality of the implementation setup may define how useful OpenAmplify turns out to be for Millward Brown or any other organization that seeks to use technologies such as the Ampiverse, today.
That's a statement that is probably true of most of the new technologies that are quickly evolving to deal with massive amounts of "Big Data" that is being generated by up to 1/6th of humanity that is now online, and is both consuming content and producing content.
I notice organizations getting excited about the potential of new Semantic Platforms (often, lets be frank about it, it's to attract new clients - or get their current clients excited about proposed new projects); I wonder if they have any idea of what it takes to really set them up well enough to make the technology useful?
I see some buzz in so far as the new technology can help sell marketing deliverables - I have more reservations about how well they are actually used - and the results I've seen so far, don't look all that impressive - probably more than anything else, because the setup wasn't done well in the first place.
I experienced it first hand with Crimson Hexagon, a very powerful technology yet the results too often did not live up to expectations when I worked with it (I can speak for myself here - so I'm perfectly within my rights) - it may be the setup was wrong or hastily done (working in an PR environment at the time, that did not allow for the right setups of Crimson in the first place, or it may be that people were of the understanding Crimson Hexagon was not in sync with the reality of what that platform delivers - or a little of both).
It really doesn't matter - the end result is what people see - and if the rest of the technology is dismissed, even when it should not be. That's just how it is and how people judge - and it's a strong argument against licensing the technology to the wrong people (I'm not talking about me, BTW, because of the policy of licencing out Crimson Hexagon, I got enough hands on experience to form and opinion - I doubt anything they have come up with lately would change it. However, if people get better at configuring using the platforms, perhaps results will improve enough down the line) for those who can not yet work effectively with these emerging technologies.
I also don't see many companies willing to commit to and pay for the kinds of configuration setup, custom taxonomy setup, etc, that would make these technologies really useful. On the other hand, I can see the need for the platform developers such as Crimson Hexagon and OpenAmplify, to try to get marketers to use their platforms more widely. Certainly implementations like this one at Poblash.org are interesting though I don't know what to do with it (the information).
A more interesting implementation of OpenAmplify is around the London Olympics, but I'm still struggling with what to do with the information.
It's an art to figure out how to take information collected off the web and decide it's uses and map it in an interface that actually allows one to act on it. I don't know if we're there, yet.
We need to think more carefully when we employ technologies like Crimson Hexagon and OpenAmplify, the setup really is crucial as well as the concept of what the results are supposed to convey.Right now there is so much experimentation going on, to see "what sticks" that it many be difficult to know just what time is needed to set these platforms up effectively; I know, I wrote a book on it.
I wonder if the attraction to this type of technology is more about an "aura" it lends to firms that start these projects then the actual results achieved - I suspect much of the talk about the potential of semantic platforms is still full of marketing hype - the potential results sound exciting - but the actual results aren't yet (exciting) enough to get all worked up about them. But they could be, if the best setups were done with more thought and time, they would be.
The good point here is that semantic tools are evolving quickly, but based on what I'm seeing (and look, I don't see every thing, so I'm happy to be corrected on this point) these technologies are far from being ready for prime time in current marketing environments; a lot of customization is needed to make it work well.
I think shortcuts to a customized taxonomy will probably not work well enough for what most people are expecting, yet. One short cut available today is within Radian6 Insights; the OpenAmplify/Radian6 Insight offers the business user generalized taxonomies applied to a River of News with a Topic Profile. You can not build your own taxonomy in OpenAmplify today, and then run it in Radian6 Insights, a real pity as that would have made OpenAmplify + Radian6 more powerful.
The idea of OpenAmplify+Radian6 is sexy, interesting and even intriguing - but lets look at the results and the reader can judge for themselves.
Just to be fair, while building a taxonomy in the Ampliverse may or may not be easy when I tried downloading the Ampliverse builder tool and didn't have much luck getting it to work - and so I can't say for sure how hard building taxonomies really are. I do know building taxonomies is a fairly specialized skill - and some taxonomies are harder to build than others, I suspect.
Here's the current parts of OpenAmplify generic taxonomy that are used with Radian6 Insights - there's no way yet to customize it with your own custom hand built taxonomy - my guess is the best thing you can do in this case is really make sure the data sources in Radian6 are cleaned up to begin with, and the only example I have to work with is the SXSW Topic Profile that has OpenAmplify activated on it - I have it via my Influence account at Radian6 - something the company has be very generous with to me, and I'm very grateful for it.
Lets look at the CRM Segments - since that seems to the star jewel of the entire set of OpenAmplify insights - the thing that sets it apart the most, it's focus on "intent" and "behavioral" activity. Here's a section from OpenAmplify product page for Radian6 Insights.
The 'CRM Segments' Insight
The CRM Segments analysis is unique and itself comprised of 5 types: Comment, Engage, Support, Contain and Inquiry. By analyzing each of the comments in your River of News, OpenAmplify is able to classify each one in CRM workflow terms. We can tell the difference between the comments that are happy and asking for help... SUPPORT... and those comments that annoyed and tell everyone about it... CONTAIN. These classifications can be a super-useful way to vector into your River of News to gain insight. Here are examples of each type of data:
Now, let's look at the actual results from the SxSW profile in Radian6 Insights the RESCUE CRM Segment is intended to show online mentions where help is being asked for:
Unfortunately, the data verbatim doesn't support that view of the data- Radian6-OpenAmplify - RESCUE CRM Segment of SxSW. Perhaps the best we could say, in this setup (SxSW) that "RESCUE" means anyone or anything that needs help - regardless of our ability to act on it (or if it's even related to SxSW).
I can argue that if this was a cleaned up topic profile, if I had vetted my sources - say I was investigating a disease and only had certain forums and influencers that Radian6 captured - and I found people who were complaining and needed help using OpenAmplify - maybe, just maybe, this information from OpenAmplify might have really made my day.
I'll even go even further and say that the setup here is more important than the semantic capabilities - a wide open topic profile of SxSW may be too expansive data set to use with OpenAmplify and get much out of it.
Does the Engage CRM segment represent fans of SxSW? Radian6-OpenAmplify - ENGAGE CRM Segment of SxSW
Probably the Engage (Fan) CRM segmentation is a closer fit - but I think there is much more power here that could have been achieved if basic configuration had been set up in a way to support market research (custom vetting sources, identified influencers beforehand) instead of PR (stick a more or less un configured listening process on SxSW and let OpenAmplify struggle to get nuggets of meaning out of it).
But until Marketers grow up - until they get away from marketing hype and deal with Analytics reality - I doubt the hype will ever come close to the reality.
Radian6 Insights promises to use the power of OpenAmplify without any technical setup. Is that realistic? Maybe we need to sell consulting with these setups to guide companies in the best use of the data - just giving them OpenAmplify is probably not going to work out too well. Perhaps in a few years the situation will be different - but not today.
In the end, when these technologies don't preform, the marketer or PR firm may need to look in the mirror.