Back in August I wrote about some data I had collected on the Old Spice Twitter account and now, six months of data later, it seems like a good time to hang up this project and move on to something else.
Before I do that though, I thought it would be good to take a look at what happened, think a little about what I would have done differently, and finally to publish the data I collected.
What Happened
There were a few things that prompted me to start this project. First was shear level of attention the project got. I can’t think of another brand effort on Twitter that got the ad industry talking more than this one. Second was the lack of data available on the project. While I’m not an analyst of any sort, as a strategist, data matters a huge amount to me and at the time I started this, baseless blog conjecture was the only form of data available. While my data won’t tell us anything about sales, my feeling remains that some insight is better than none. Finally, I’ve been doing this long enough to know that the ad industry is great at remembering the names of successful projects, but a lot less good at remember how or why those projects worked (subservient chicken, I’m looking at you). My goal with this is to provide some level of historical context for this project so that people can look back understand it at hopefully at least a slightly deeper level.
So what insights can we gather from 6 months of data?
For the most part, I think what I said in August still stands up.
“Much of the conventional wisdom around brands on the web these days centers on the notions of communication and reciprocity. The idea here is that if a brand wants to be successful within the context of the “social web” they’ll need to act a lot more like people and a lot less like companies. But looking at the Old Spice campaign - I have to question some of that.
It’s worth noting that the Old Spice account follows back less 1% of the people that followed them. Also, their rate of communication is about .8 tweets per day. At the same they have about 1% daily increase in followers - about 1,000 per day. Basically - @oldspice was looking a lot like a celebrity account: lots of followers, very little following. This had me wondering if people were following Old Spice the brand, or Isaiah Mustafa, the spokesman? Further confusing the issue though is that unlike those accounts, there isn’t much human connection coming through the account. It’s mostly humorous non-sequitors, and even then, there’s not much of that being produced.
In fact - nearly the entire catalog of bi-directional communication, supposedly the point of brands in the social space, happened in a very short window right before the end of the campaign. This was the time when Wieden was staged their famous video twitter responses.”
I ended that post thinking two things:
1) People who followed the account we’re connecting with Mustafa, not Old Spice
2) That either the conventional wisdom how digital marketing works was wrong, or that the Old Spice campaign was actually a traditional T.V. campaign that happened to be on Twitter.
Since then, a couple things have happened that I think bolster point one and begin to point to answer for point two.
From June 22, when I started this, to September 8, the Old Spice account went from about 91,000 followers to almost 119,000. A net gain of about 28,000 followers in about two and half months. From September 9 to January 23 the account went from about 119,000 to about 119,840. A net gain of about 840 followers in about four and half months. Although the rate of tweeting stayed at about 1 per day, and the tone of the campaign remained consistent, people were clearly no longer interested in following it.

What happened on September 9? That was the day WK launched their new campaign for Old Spice with Ray Lewis replacing Isaiah Mustafa.
Back in August I closed with a couple questions
“…was the Old Spice campaign one of the best social media/web/interactive campaigns ever, or, was it actually the perfect example of what a post-web T.V./broadcast/traditional campaign should be?
If it’s the former, than I think we in this industry need to reexamine our canon of what makes great digital advertising - because we seem to have gotten a lot wrong.
If it’s the later, than I wonder if this isn’t an accidental (or intentional?) example of just how effective the internet and the web have been in totally blurring the lines where content lives and instead leaving us to focus entirely on the nature of the content - in this case, traditional “lean-back” content using Twitter as a distribution channel.”
As I’ve watched this campaign go forward, I feel more and more certain that the Mustafa era of Old Spice was in fact the 2010 version of how traditional advertising should be done: tightly integrated across multiple channels, but using the mechanics of traditional broadcast advertising.
Given WK’s mastery of this format, it’s not that surprising.
All the activity in the digital space, and especially on Twitter, was almost completely tied to television. So much so that the only real activity that happened on Twitter was when it was used to direct people to other commercials. Commercials made in near real-time, but commercials. Whether intentional or not, the Twitter account never had a life of its own outside the T.V. spots and the character they created.
So what to take away from this? Well, without access to WK’s long-term strategy for the brand, without clean sales data, and without budget insights, it would be hard to draw any meaningful conclusion about overall success based on a handful of Twitter stats. But I think it does serve as a brick in a larger understanding of the nature of brands online.
It’s not at all a shock that if your goal is to create an online footing for your T.V. campaign, this would seem to be a great model. If, however, your goal is to create an active community outside your T.V. advertising, it would seem the conventional wisdom is at least more right: you’re going to have to create a uniquely compelling and ongoing reason for people to join.
Just as with T.V., single spikes in activity online seem to only create single spikes in interest. The important thing is understanding what you’re trying to achieve - no single approach is good for everything. If your goals are the sort of goals advertising is good for, than I think it could be worth looking at how WK conducted this campaign.
If this campaign reenforced anything for me it’s that in the end, advertising is advertising, no matter where you put it. The web is an amorphous medium, capable of looking like a lot of things - though some better than others. What it doesn’t do is take a set of ingredients and make them something different. If you pour advertising ingredients into the web, advertising comes out.
What Would I Have Done Differently?
Well, a lot.
First of all, I wish I would have tracked a couple other accounts to compare against. For example, back in August I made the comment that Old Spice account felt a like a celebrity account to me - I wish I had been tracking at least one celebrity account to see if it acted the same way. Also, I wish I had been tracking another CPG-type brand just to set a baseline. Finally, I wish I would have tracked my own account over this timeline, again as a baseline.
Also in hindsight I just wish that I had more data. If I had to do it again I would have at least tried to keep track of @ replies coming from and going to the Old Spice account. I want to have a better sense of what the conversation, or lack-there-of looks like. Something I have in my notes, but didn’t track well enough to include it in the final data, were the kinds of tweets the account sent out. At some point I noticed the account started sending out more DR-type tweets and it would have been nice to be able to see what effect those had.
At the same, as much as I would love to have this data, tracking Old Spice data isn’t my full time job and making the process of data collection easy is part of the reason this happened at all. I could have set up a Radian 6 or ScoutLabs account and gotten everything I wanted and more, but I wonder if I would have been able to keep it up long-term. Unlike the work we do here at Fight for our own clients, this one was purely a side project for me.
What’s Next?
So that’s the end of this project. I already have a couple ideas for followup projects and hopefully I’ll be able to take what I learned from this experiment and make those better. I have to admit - it’s hard to have 6 months of work into something and then start over knowing that it will be months before I have a useful base of data again. On the other hand, I think the marketing industry as a whole, and advertising in particular, is overly obsessed with “today.” We too often lose the the insights that can only be gained by watching how “today” turns into “tomorrow” and then “6 months ago.” Sometimes the only way to uncover the fundamental truth of something is to be able stand back, sometimes miles back, and look at the broader ways things relate to each other. It’s understanding those truths that drives me as a strategist, drives everyone at Fight, and ultimately is what I think results in the sort of culture shaping work I think we all want to make.
The Data
This zip file contains my original Numbers file, and an exported Excel file. The Numbers file has all my charts formatted properly but I think the Excel file should be serviceable too. I marked in purple the date (9/9/10) as the date Ray Lewis was introduced to the campaign. Marked in orange are 3 days that show what seem to be odd changes in the number of tweets shown for the account. I’m not sure what’s going on here - it could be something totally normal that I’m just unaware of, or it could be some system error on Twitters part. I’m just not sure.
This data is, for whatever it’s worth, free to take and do whatever you want with. If you do something interesting with it, I’d love to hear about it.