Figure A: The results of a drunken flamboyant slug race across a sheet of graph paper

Regular readers may have guessed from recent posts that I have a Twitter obsession. Let me tell you why: for the first time, Twitter has endowed us with the ability to query and analyze conversations, opinions and proclamations. In real-time. For free.

I don’t feel like we’ve fully grasped this yet. People all over the world are talking to one another, and we have the capability to filter, aggregate and rummage through this data, without the damaging bias that is introduced when people know that you’re analyzing their thoughts.

How could we use this awesome resource as content strategists? Here’s one idea to get you started.

The above graph was constructed from two weeks of San Francisco tweets. It shows how the use of certain words changes throughout the day, adjusted to take account of the typical daily rise and fall of tweet volume.

There is a clear peak in use of the word “today” at 8am-9am, and again at 1pm, possibly reflecting on the morning’s achievements and grievances. “Work” appears to be a constant worry throughout the day. Thoughts of “tonight” and “tomorrow” gradually increase from 8am, with the former peaking at 7pm-8pm, and the latter at 10pm-11pm. “Yesterday” is the least mentioned, and is given slightly more thought from 8am to around 2pm, dropping just as talk of tomorrow ramps up.

Interesting stuff, but perhaps not particularly practical to content strategy.

The point is that we can easily conduct similar analysis for the audience of a website. What do they talk about? When do they talk about it? Who do they talk about it with? What kind of language do they use?

The data can inform website nomenclature that is better suited to the audience, and is more likely to surface against their search engine queries. Content can be published at times that slot into their schedule. Trends and opinions can constantly influence the subject-matter of content production, not necessarily reactively, but at least as part of a proactive ongoing assessment.

We’ve become accustomed to the traditional – albeit valuable – methods of market research, user interviews, persona construction, user scenarios and focus groups. All these techniques take snapshots of an audience that is more in-flux than ever. Twitter gives us a tool to progressively monitor and adapt our audience models, at low cost. Even without the powerful nuances of user interviews (e.g. the all important “Why did you say/do that?“), we can at least use the data to detect shifts in audience behaviour that can trigger a manual re-assessment of who we’re dealing with and what they need from us.

It’s just an initial idea, and as usual, we’d love to hear your thoughts on the matter. Ah, but actually, you don’t have to tell us. We’ll spy on your Twitter reactions instead.