Exploiting Social Media – Some lessons

The past 12 months have been a major learning experience as I transition from my recent social media startup.  After a hiatus of nearly a year from blogging,  I wanted to summarize some key lessons as I rode the downslide of the first social media wave. These include:

  • Marketers are still trying to figure out does social media work (in a reliable manner compared to other online channels). Selling tools and me-too products is just not enough.
  • Social media advertising is still a variant of  the good old genre – “display advertising” – call it promoted tweets, sponsored stories etc. From a marketer’s perspective, easier to do another ad copy for a channel rather than focus on costly prospect specific engagement.
  • Running social media campaigns effectively is a costly affair. You need a big Ops team to do anything sustainable.
  • A big question: Is social media advertising even worthwhile for folks on the long tail (in their category)? Is there even enough worthwhile volume of chatter for leads or otherwise?
  • Social media analytics – NLP, big data crunching etc. is quite erroneous and noisy. The Signal-to-Noise ratio (Read Nate Silver) is quite low.  To get real value, one really needs to do a lot of work. The big question is: Is this all worthwhile ?
  • Finally, mobile and social are quite conflated – as every day consumers try to figure out the modalities of what works best for them. Marketers are still grappling with the question which channel to put their spend on.

Answers to many of the above questions are anecdotal with the end justifying the means.  After processing nearly 100 Million tweets (Twitter) and statuses (FB) in specific verticals (over the past 24 months), less than 10% of these tweets have real economic value. These numbers vary by vertical but social media has long ways to go to compete with display or search as marketing channels in terms of funnel volumes. Having hand-labelled nearly 100000 tweets for our engines, I believe there are pockets of worthwhile chatter. Considering the origins of twitter (Short staccato language used by cops, emergency services, dispatch operators),  one wonders how much of “semantic” content can one embed in these tweets. Much of the these streams are awash with brand/publisher driven chatter that finding worthwhile nuggets still remains a difficult problem.

On the upside, I hope as folks learn how to use and exploit this medium as publishers, advertisers and “readers”, the quality of content (UGC or professionally done) will mature, leading to some real value creation.  With the introduction of video, audio and other content forms, newer applications of this channel remain to be discovered. Newer forms of engagement, content types and inter-platform integrations portend richer consumer experiences in the near future.

Nearing Half-time in 2012 – Checking on tech trends

At the beginning of 2012, I’d put a list of potential tech trends. Checking in on them five+ months later, this is what we see:

  • The revenge of the “publishers”: Does not look like publishers have figured anything out yet. Quite swamped with the plethora of devices, tools and channels!
  • Integrated advertising: There is an improved interest in this issue. Other than large investments in DMP’s, the how-to is still an open issue
  • New ad creative formats: Have not seen anything exciting yet. GM demanded some new formats from FB but nothing new on the horizon!
  • Changes in media consumption habits of consumers:Lots of anecdotal research. However, little stable behavior. With tablets entering the mix of devices at scale, this issue got more complicated!
  • Global sourcing of tech talent: Lots of startups on techcrunch seem to be focusing on this, but validating quality is an issue not yet solved. Global sourcing is till very process/KPO oriented rather than product development oriented.
  • The integrated open source toolkit: Nothing yet. More divergence rather convergence. As a starting step in the right direction, check out Clojure with Datomic !. More maturity and robustness in the different opensource toolkits and languages.
  • Enterprise apps revisited with cloud and social media: Some high profile acquisitions. Only the future will show these are being used within an enterprise.
  • Integrated consumer view of Amazon, E-Bay, Google, Facebook,Yahoo, Twitter: Not seen one yet..but lockerz and the like seem to be building one around a social id.

Nothing majorly new other than the ebb and flow around the FB platform and the patent tangles.  Open question to follow – Is “social”  desktop centric or mobile centric ?

Looking Ahead – Interesting tech-oriented themes for 2012

The past quarter has been quite hectic as I have been driving the technology efforts at my startup, Salorix, in the social media advertising space. As we ramp up our development, I have been reviewing a wide range of  industry trends and technologies, and a few key themes stand out.  They are in (no priority order):

  • The revenge of the “publishers”: Content generators such as newspaper/magazine/media (TV, Radio) and book publishers, movie houses, game studios are revisiting their content monetization models. Key question: Do they really need intermediaries such as search engines, bookmarking sites, microblogging and social networking platforms to drive traffic to their content? Maybe the focus will turn to retaining/engaging their loyal customers.
  • Integrated advertising: As the interplay between online/digital and offline campaigns gets complex, advertisers and their agencies will revisit their campaign strategies. Key question: How should I manage my ad budget effectively so that campaigns leverage each other across channels for total effectiveness rather than piecemeal metrics by channel ? Maybe the focus will turn onto an integrated view of the traditional sales funnel.
  • New ad creative formats: As social media advertising ramps up, advertisers are considering ad formats that drive “engagement”. Key question: What can one do that is different from a user/visitor clicking and getting to a landing page? Can we get  them to participate in a game? respond to a survey? collaboratively watch a video ? Ad formats that coordinate “engagement” in various ways may be next on the horizon.
  • Changes in media consumption habits of consumers: With the plethora of devices and media, consumers are opting to choose different kinds of devices for consuming different kinds of media. One may catch a quick summary of a primetime show on their mobile,  view a review of the cast on their tablet, watch older episodes on their laptop (catching up with the plot) and finally watch the show on their new LED TV.  Key question:How does one learn what consumers will do/engage with the “content” you produce ? Where will they engage best with your brand message in what kind of format? Figuring this out (in a dynamic realtime manner) may be the next gold mine.
  • Global sourcing of tech talent: The basic building blocks of building next gen software systems are getting standardized. Folks worldwide have access to these tools and can learn them on their own free time and become adept at it. To bootstrap your next startup, may be all you may need is a good definition of your product and your tech team will be built from experts worldwide! Key question: How do I go about finding talent and building my best teams virtually?
  • The integrated open source toolkit: In a manner similar to LAMP/MAMP technologies, the other building blocks such as DBs, Search and Messaging may be integrated into a super toolkit for building the next generation system. These also may integrate consistently with the broken world of APIs ad newfangled storage services (such as DropBox and the like). Key question: Who is cobbling together this next gen toolkit?
  • Enterprise apps revisited with cloud and social media: Monolithic enterprise apps of the past decade are giving way to a new generation of systems that are cloud oriented, remove unwarranted functionality by enabling SAAS deployment, and tap into unstructured organizational knowledge via social media. However, it is unclear how enterprises should adapt to this trend. Key question: How do I migrate? What about my current org structure? What about legacy? What about all the investments I made in .NET ?
  • Integrated consumer view of Amazon, E-Bay, Google, Facebook,Yahoo, Twitter: There are only so many things that the casual web savvy user does on the web on a daily basis – find info, browse info, buy, share info/opinion, engage socially, consume entertainment (music/video), possibly create. Why do I need 5-10 apps to manage  these in my daily life. Key question: Can one get a one-stop shop app for all services that I consume – that pulls all of this activity modes into one utility.- best of breed without losing my privacy across all my devices ?

Even if one of these trends pans out in the coming year, that will be cool.

Computational advertising – where are we?

Computational advertising has grown by leaps and bounds over the past decade. As Digital marketing techniques and platforms grow, there are  a number of new open problems that are on the horizon spanning a range of fields. I briefly summarize them here.

  • Audience Behavior – How much more do we know about the “audience” in  a reliable manner before the days of online advertising? Can we know more without violating the bounds of privacy?
  • Content Analysis – How do we integrate content analysis across text, audio, video, images to get a reliable view of the audience and brand perception?
  • Integrated Channel Management – How do we coordinate campaigns across email, web/SEO, SEM/PPC, social, mobile, video, audio, radio (online/offline), print and TV?
  • Active vs Passive Engagement – Transition from a one-way mass communication paradigm to a 2-way – one-on-one conversation puts immense load on both the marketing infrastructure and the individual audience member. Is this the way forward? or is there a happy medium?

With all the technologies in play, the advertising “workflow” is still open-loop. Marketers message people (in a number of ways – push or pull), and the audience responds either online or offline, near instantaneously or in a delayed manner co-located with the message or not. All the data available through this workflow is noisy and fragmented either by design or by operation. However, the connection between audience response and the initial marketing strategy is made in a heuristic manner. Filling in this gap in a reasonable manner would be a worthy challenge for different players in the eco-system.