Based upon the 30 billion public social activities we are delivering each month, I can confidently state that the largest companies in the world are in process of figuring out how they can incorporate public social conversations into their daily business operations. From supply chain management to PR crisis management, the role of public social data in the enterprise is no longer framed around the question of “why does this data matter?” That said, we are still in the very early stages of corporate adoption and there are plenty of unanswered questions for most companies. Once enterprises get past asking why this data is important, the next obvious question to address is: which public social data is best for performing business analysis and decision-making? Facebook? Twitter? Google+ (s GOOG)?, WordPress (see disclosure below)?
We’ve been trying to help our customers answer the “which data” question for the last couple of years by providing anecdotal insights via examples we’ve seen across the industry. Although this approach has provided reasonable guidance, we’ve recognized for a while that there was an opportunity to take a more systematic, research-based approach to answering the question. This summer, we hired a Data Scientist to help us analyze and understand the data needs across various business use cases. The early results suggest that there are at least two key attributes that play a role in answering the “which data” question — reaction time and depth.
From the launch of the next gadget to the death of Osama Bin Laden, significant portions of social conversations are based around events. By analyzing social conversations around specific events, results quickly emerge illustrating that some social networks have faster reaction times to events than others. For example, we looked at the social conversations happening around Netflix on October 25th, which was the day after Netflix announced their most recent earnings.
As you would probably guess, Twitter reacted quickly to the news that Netflix opened down significantly. You can see from the chart that peak Twitter conversation happened within moments of the market opening and quickly trailed off. Studying other conversational data sources such as WordPress it is clear that the overall reaction time was slower as compared to Twitter and the area of the conversation curve was much greater. As you would expect, the reaction time for blog comments from WordPress significantly trailed the blogs themselves. It is interesting to note that peak Twitter activity corresponded to the opening bell where WordPress comments peaked well after the market closed. I’m highlighting a very compressed event here, but we’ve studied enough events to see that these curves hold for slower developing events too such as political elections.
One other aspect of social sources as they relate to reaction times is whether an event was an expected event or an unexpected event. The reaction time curves behave differently between a hurricane (expected natural disaster) and an earthquake (unexpected natural disaster), but I’ll have to save the details of this for a future blog post.
The amount of content that can be analyzed for insights across public social conversations varies significantly by source. One of the key benefits of a source like Twitter is that the reaction times are fast which is great for business use cases requiring early signal detection. However, the amount of content you can analyze for deeper insights on a single tweet payload itself is relatively concise. On the flip side, sources like blogs, forums, and videos are more likely to be rich in opinion, sentiment, and engagement. In our Netflix example, Tweets such as “NFLX gaps down 35% on open!” tell you what happened, while blog posts like “How Netflix Lost 800,000 Members, and Good Will” have the deeper analysis to tell you why something happened. Measuring depth is non-trivial, but there are lots of ways to get a feel for the depth of a network. For example, one simple way to get a reasonable level of accuracy is to just look at the byte size of the various social activities. Another way to gauge depth is to look at the area under the reaction time curve. The bigger the area, the more likely you are to get a wide range of views and perspectives.
Mixing the perfect social cocktail
Which public social data is best for performing business analysis and decision-making? The answer: it depends upon the business use case.
PR crisis management is typically very dependent on ultra fast reaction times so a source like Twitter can be great for early signal detection of a potential problem. The same company might also be analyzing different public social conversations to inform their product development roadmap. Millisecond reaction time for product development use cases isn’t typically a necessity. In fact, a company may prefer in-depth analysis of historical conversation data over realtime conversations in order to make decisions about future products. A far more important component for product development is the insight that can be gained from analyzing deep social conversations, opinions, etc. One thing is clear from our early exploration; very few business use cases can be served completely by a single source of public social data.
Lots more work to do
This is just the beginning. We are continuing to explore the role of different social data sources across different use cases. Reaction time and depth are two important components, but we also recognize there are plenty of other variables to consider. As we discover more we will continue to share our findings. Our goal is to help businesses understand the appropriate social cocktail for the use cases they need to support. Stay tuned …
Chris Moody is the President and COO at Gnip, a realtime social media data delivery company. He serves on the National Technical Advisory Board of Year Up, is an advisor to several technology startups, and is an active TechStars mentor.
Disclosure: WordPress.com is backed by True Ventures, a venture capital firm that is an investor in the parent company of this blog, Giga Omni Media. Om Malik, founder of Giga Omni Media, is also a venture partner at True.