I promised Synthesio’s Community Manager, Michelle Chmielewski, a review Synthesio’s new dashboards for close to two months; looking at a live demo and a link to a brochure that explains what the new features / changes are. Having not worked with this platform hands on before I found touching the demo and it’s data, refreshing. In case anyone is wondering, Synthesio is a social media monitoring platform with it’s headquarters in Paris, France.
Everything I’m talking about you can go and see for yourself (go ahead, try it).
Note: I have no official affiliation with Synthesio but I know many of the people work there. I have seen the platform evolve this year and it appears they’re growing fast and are definitely a platform to watch and consider using.
First off, Influence Mapping has been fairly poor for most of the platforms I’ve looked at but Synthesio has done a pretty decent job and added some new features that are generally lacking elsewhere including filtering on industry segment, topic, tone and location (country). (Note: some this functionality is present in the other platforms to some extent but it often requires changing the topic profile in Radian6 or additional filters in Sysomos, etc).
Synthesio Influencer Dashboard with Scope=General News, Magazine, Topic=Prius, Tone=advocates, Country=France and Period=8/27/09 to 11/26/09
The influence dashboard is immediately more useful than what Radian6 or Sysomos currently provide due to it’s intuitive feel and grasp of the data that is easy to get hold of, and you can look backwards and visualize the sources of links (influence) that connect to any of the influencers that show up in the dashboard (but, I can hear Sheldon at Sysomos saying the same thing can be done in MAP as well, and that’s true, it can).
To be fair, Sysomos has a similar visualization (not as pretty though) while Radian6 lacks this functionality altogether. Synthesio has it’s own rank and it’s calculated much as Sysomos’s is.
Numerous criteria for influence are available including :
- Number of verbatim, Score of influence (Synthesio Rank) between 0 and 10
- Score of overall influence (combination of the 2 preceding factors)
- Audience in number of unique visitors per month (data from Google AdPlanner)
- Average sentiment of the verbatim
- Alexa Rank
- Google PageRank
- Number of inbound links
- Number of verbatim published by the source over the past month Clicking on an influential site will bring you to a detailed page showing the verbatim for that site as well as the site’s profile (audience, topic(s), readership, logo, etc.).
So I see the influence dashboard with it’s multiple filters as an evolution of the platform.
I like to consider why someone pulls data in the first place and what kind of questions are they trying to answer? What actions can the client or business take or want to take?
Take this use case (refer to the image above): Suppose I wanted to find influencers for Prius car on Forums that were located in the US, (Synthesio’s Influence dashboard would get me to about the same place Radian6 does with a Topic Analysis widget breaking down by Region, etc) down to a select list of forums. Actually, Synthesio’s influence buttons on the left of the screen above look like a good way to get at the data (and to be fair, most of the data still is present in many of the other platforms I use, but it may be harder to get to it and more importantly, doing some correlation analysis on the data may be easier or harder, depending on a platform organizes information. In the case of a group of Forum influencers, the buttons on the left are of no real help, but then, none of the other platforms is of much help with this use case.
- then the real work would begin as I’d manually have to go into each forum myself (or have someone else do it, serious intern time!) and decide who was an influencer that I wanted to reach out to (via # of forum posts, frequency of posting, on topic mentions for the subject I’m searching for).
These things Synthesio tries to provide as well, but due to the date restrictions in the demo I wasn’t able to bring up anything and the image below is from their brochure.
Facebook Influencers (individual users)
List of the most influential Facebook users for a given topic, with the possibility of filtering by number of contributions or by score of influence(which takes into account the influence of each contribution). Exporting the data to Excel is possible as well as demographic stats about the number of users, sex,country, language.
Facebook Influencers (pages)
A list of the most influential Facebook pages or groups for a given topic, with the possibility of filtering by number of fans (for a page) or the
number of members (for a group), or by number of contributions.Exporting the data to Excel is possible as well as a visualization over time of the number of fans and contributions.
Synthesio does have the influencer, twitter, Facebook and facebook person icons on the left of the influencer dashboard but they didn’t work for me in the demo for what I was trying to do (it’s a demo, after all).
However, if it did work (it probably does for a fully populated dashboard for a paying client) it would take me much closer to where I want to go and where I think most clients want to go, the individual. And, it’s also where PeekYou‘s value would shine if merged as you get a fuller profile of the individuals by demographics and psycho-graphics based on current public data such as zip-code which can then be correlated to income levels, etc.
There’s also nice visualizations of Twitter influence (probably by number of followers I’m guessing) that are sprinkled within the interface, they may not work for this demo but I’m sure they come in handy for paying clients.
My only caveat is the data set that is being used will be pre-populated and set up by Synthesio for the client and may not allow for a full exploration of any subject for any query that might come up in an agency model (I have to check back with Michelle Chmielewski on that).
For example, I want to do research on anything I want, I can just use Sysomos MAP, write a query and pull whatever I want from their database, but in Synthesio’s case, they probably want the query first, refine it, pull the data, clean it, and then populate the database. No? If that’s the case, you need to plan what your going to research and report on, and pay for, beforehand.
You may not yet be able to use a platform like Synthesio to get influencers for bubble gum unless they set up the bubble gum database for you first (and additional charge, of course). It’s not as if Synthesio gives you access to the whole wide world and you can put any query and find any set of influencers you want, ad-hoc, like you can do search in Google.
I just point it out because platforms are designed to for different purposes, my guess is Synthesio is best set up to be useful for a customer who knows what they want well ahead of time as opposed to someone, say like myself, that would want to use it for “general research” on say, bubble gum, or say …. diet soda. Unless I get them to set me up for that, am i going to be able to use the tool that way? My guess is not, but Michelle, tell me otherwise.
There’s several additional nice visualizations that Synthesio provides, I won’t cover them all here but the brochure that you can download covers them. The breakdown by Mainstream vs. Social Media is nice and Sentiment Analysis by Verbatim also looks interesting.
Sentiment breakdown by topic and topic evolutions seem pretty standard, and you can get the same thing from Alterian/SM2/Techrigy and several other platforms so I won’t really spend any time on them here other than just mentioning them, that Synthesio has these features because clients need them for their dashboards.
The Topic Mapping seems to be an improvement over what Crimson Hexagon offers in their Opinion Monitor, because this data is already cleaned by Synthesio while Crimson Hexagon’s diagram is generated by algorithms and a sampling of a small set of documents are used, and just the tagging on those documents are used by CH. Synthesio’s topic breakdown is probably superior in that it is based on the actual analyzed data where as Crimson Hexagon’s is not (and I have brought that up to Crimson Hexagon a few times and they’re working on it).
The topic mapping is just another way to find the data you want, but it’s a good one. Compare this to Crimson Hexagon’s topic mapping (I don’t have an image online I can draw from but the finished product looks like this, below). The finished result based on machine learning is only as good as the analyst that trains Crimson Hexagon and depends if there’s actually enough data. Synthesio’s approach is much different but there are pluses and minuses to both (that will be covered as well in my book)
BTW, an expanded version of this material will go into my book and if readers give me feedback on what they see in Synthesio, I’ll address it in the book (as well as perhaps, writing an additional post to address features that were left out or need clarification).
