A Sale of Two Twitters

On Monday I attended the Media 140 Conference. Owing to a confluence of adverse railway related issues, I arrived late (Am reconsidering my opposition to fascism, I think Focussed Fascism - on the trains - would be a Good Thing) so missed the first morning session (but Adam Tinworth has blogged the whole day starting here). The aim of the day was to explore how microblog systems (ie Twitter) work in a more commercial arena for brands, ie "Everything a brand needs to know about twitter & real-time social media". This meant, on the day, focus was mainly on sales, marketing, PR and the occasional dive into customer service rather than any more operational uses.

At any rate, these are the notes I took on the day.:


Quirk's Nic Ray and Unilever's Noam Buchalter on Crowdsourcing

Crowdsourcing is not new, but new technology helps - cheaper, more people, faster, do more things. It attacks the traditional Ad agency model's economics as:

- there are more creatives outside
- it is lower cost and risk plus faster turnaround
- potentially gives better consumer insight, and often earlier
- drives better consumer exposure
- has high PR exposure at moment (buzz)

However, its not without its challenges:

- sifting ideas takes up a lot of time and effort
- manage the contributing community expectation so they remain "customers" (and prevent backlash etc if they feel they are ignored)
- these, and other production issues still need paying of skilled people
- IP issues with 3rd party ideas
- old agency sour grapes - got a lot of negative press from people who have access to presses

Also, Crowdsourcing is not appropriate for everything, works best with:

-fairly well defined brand
-something the target market would have a view on

They went through a case study for Peperami, who changed to this model, in essence because:

- wanted new ideas from new sources as brand was getting old
- economics of outsourcing to (less expensive) users were very attractive

After the users generated the ideas, they were sifted and scored by:

- fitted client brief
- was actionable
- "wow" factor
-was produceable

What was fairly radical apparently was that the client was fully involved in the process (I must say, coming from the consulting/design side of the service industries that point amazed me - seems like there area few other issues "Olde Advertising" has to deal with still).

There was then a Panel I missed owing to Real Work Issues, but Adam Tinworth has it covered (All panellists are listed in the Media140 link above in Para 1).


Ciaran Norris of Mindshare on Listening skills....

Ciaran addressed the economics of "Social media" marketing - he noted that points not often grasped up front are that:

- Listening is very labour intensive
- Shirky - transaction cost and the creator effort is much lower - but this means there is a lot more stuff, much of which is dross (my words)
- Search is the biggest social tool still
- Buzz metric tools are not cheap - search, viral tracking etc - once you scale up.

Also today one needs to track video not just text, and there is a proliferation of listening posts

- Delicious is a good listening tool, has RSS feeds
- Tweetdeck good for listening too
-Tweetfunnel - allows multiple people to manage Twitter
-Tweetmeme tracks links spread around Twitter (Ciaran hypothesis 90% of content shared on Twitted is professional, not amateur)
-Twitterfall searches a hashtag
-Trendistic etc etc - all seem to be point tools

He also made the point that the brand may have its online personality hijacked - For example, Weight Watchers on Twitter is not expressing it's personality, its just a recipe feed - so "Tweet what you eat" has hijacked the brand personality on Twitter by being far more about peopel commnicating to lose weight

Lunch and another Real Work Break so I missed Mr Red Bull, John Beasley, and then we had.....

The Great Twitterwall Hijack

There was then quite a useful panel with real experience - or at least it could have been except for The Twitterwall. As far as I can see, the dynamic went something like this:

- Panelists stop talking when Twitterwall goes up and start reading Twitterwall and responding to it
- This encourages twitterwalling audience to become more strident in their twts, trying to be seen by panellists
- Result is random, diverted panel with not a lot of useful stuff.

That at least was my view, I'm coming to the conclusion that Twitterwalls should be shut off (or at least curated) while panels are talking and only switched on at question time, or there is little point in having the panellists. Anyway, Adam Tinworth covered this panel well (and disagrees with me re Twitterwalls....).

The Last Panel of the Day

This was quite useful as (i) it had a real user - Virgin Trains and (ii) it had Will McInness and Drew Benvie on, who I think know their stuff (as in fact did the others, I just didn't know them).

Richard Baker from Virgin Train made some interesting observations (see his blog post over here)

- brand is in hands of the people in the company. Some customers want a conversation, others just want info/answers
- to start up: no strategy, started on Twitter, just listening. People very positive - but after a while it's not enough - the service actually needs to work
- issue with the instant gratification of real time media - why should a brand respond instantly? There is a penalty if you don't, but its lousy economics sometimes
- but people are ruder to Virgin trains than to his own persona

By the way, the Grauniad's Mercedes Bunz (yes, really) covered this panel over here.

The Real Time Web & Show me money - Bernard Desarnauts, Glam Media

The Real Time Web and how to make money - two topics dear to my heart! Here is the Problem:

-Fragmenting content
- fragmenting audience long and mid tail
- fragmenting traffic
- classic web less important (I love the idea of the Classic Web)
- brand engagement higher in distributed media

The Solution - analysis tools, in this case Tinker, a sort of Real Time Walled Garden which allows brands to see in real time stuff but gives them a safe place. The main issue though is the same as with any Real Time system, ie stopping clutter. The solution - blending algorithm and curator via real time categorisation of type of source, aggregation and filter plus metadata curation eg of hashtags.

Other things coming down the road in Real Time are:

- Event ads and real time ads
- Integrate real time web data - for example real time location targeting
- Where was the money and how does it flow? Provided metrics will be key in future
- Data send to user on various sites eg Twitter depending on what they are on, will see integration to Twitter, Friendfeed, YouTube.

Why will these approaches work? Money. Early indications are that it works 5x better than other Ad systems

- 17 vs 2.7 % interaction rate
- 15 vs 2 seconds engagement times
- 28 average tweets into Ad unit

As to how they are filtering, for eg how do you screen and remove noisy inanity - you can't yet, so they are scrubbing repetition

Dave McCandless - InformationIsBeautiful

Dave's view is that information visualization was "a new trend" (consternation in the back row as we all tried to work out if it was 10 or 15 years ago that it was a hot new trend). Anyway, Dave's view is that to to sell on social media, you need to give something - and "interestingness" is great currency (if you are not a sleb, that is?). At any rate, Dave provided some brilliant graphs and data visualizations, one of which I've put up. (By the way, the brevity of this precis is more than made up by the graphics he showed. You can get (more here)- a picture, as they say, is worth a 1,000 words)

Utku Can Akyuz - MintDigital

Utku put up the Hemlock Open Source middleware applications, which seems mainly optimised for making game type environments. Applicatios mentioned were:

- Games: Real time open source tools for eg bamzooki as game
- Filtering and graphics
- 2 Screen behaviour TV plus Lsptop integration - made as game
- Data from the ads to interact with, or for eg pulling data from shows for Ad placement next to TV

Mark Rock - AudioBoo

As the last speaker, Mark did the sensible thing and made his talk short and sharp. Key points were:

- Twitter made SMS social, trying to do same with audio at audioboo (and video in future maybe?)
- Podcasting was too complex, turned people off
- AudioBoo cannot be edited, deliberately trying to keep it fresh and simple

So what's with the headline - a sale of two Twitters - I hear you ask?. Well, the interesting stuff is often what is said in the discussion at lunch etc, and there were two strands of polarisation I picked up - firstly, the view that not a lot new was being said, but what was being said was to new people (In this respect I thought that it varied, the individual speakers were quite perceptive but sometimes the panels struggled to get insights out, as it were).

The second polarisation was a philosophical difference about how Twitter is used by Brands, 2 "tribes" in this space as it were. These tribes are:

(i) The "Sell, Sell, Sell" fraternity - the view is that some people see this as just another channel to flog stuff on, and stuff the other users' experience. The channel has a higher response rate still, so make hay while the sun shines and the devil takes the hindmost. (The "Tragedy of the Commoners" approach)

(ii) The "Social Sell" fraternity who believe that you do actually have to act differently on this medium, be part of a conversation (ie the "Its a Sewer" fraternity - what you get out of it depends on what you put into it)

There was also a feeling that the "newbies" are predominantly of the former persuasion, the old hands of the latter.

Thoughts of Which Twitter will win out?

(Update - as always, kudos to the organisers, I know how much work is involved in things like this)