FEATURE8 October 2012
FEATURE8 October 2012
Twitter’s foray into research raised eyebrows last week. We asked industry experts if the social network is pursuing the right approach to measuring ad effectiveness.
Brand strategy VP Joel Lunenfeld said: “We’re giving brands the ability to deliver and measure the impact of mobile and traditional desktop campaigns through these surveys.”
Coke’s Johan Houben was intrigued by the move, but what do research experts make of it?
The screenshots for Twitter surveys look very promising and prove that this is not an old fashioned text survey re-imagined. They are going well beyond 140 characters.
I am in favour of any method of research that fits into consumers’ lives and allows easy participation. In addition, being able to map stated attitudes from a survey against observed behaviours and influence on Twitter would deliver great insights for brands – assuming that @TwitterSurveys is able to deliver that sort of feature down the line.
But that said, there are a few issues with tapping straight into the timeline. For example, take @HonestToddler. Should the anonymous adult writing the tweets take a TwitterSurvey “in character”? Or as themselves? The obvious answer to those of us in the research world is as themselves, but the line is likely to be somewhat blurred to the person who is creating content for a character or brand behind a Twitter handle. Although this is a fairly extreme example, we can’t ignore the fact that many people have more than one Twitter identity and use Twitter for some very unusual reasons.
Dear Twitter, welcome to the survey world. Given that Facebook and LinkedIn arrived a long time ago, you’re late to the survey game. Regardless, the technical capability of launching questions and receiving answers does not a robust technique make. Any survey tool that collects data from a unique and siloed population is guaranteed to build biases into the results.
We know for a fact that opinions gathered from one social media site can be massively different, even opposite, from those gathered from other sites and that makes for bad brand decisions. Users of the Twitter tool will have to ensure that they incorporate Twitter surveys into a wider and more complete research program if they want to ensure any level of validity and generalisability.
I’m betting the Twitter team is already on that road. Let’s wait and see.
It’s difficult to say right now if Twitter needs to do more work on this proposition but I imagine these surveys will be targeted at people with known exposure to advertisements on Twitter. As we all know there is a great deal of pressure on publishers and digital media to prove their value to the media buyer.
The only issue I see with other media outlets following in Twitter’s footsteps is that they will not have a tool to reach out to their members ‘on the fly’ post-exposure. However, by building intertwining research panels and tracking tools on their audience, publishers should be able to compete on an equal footing with Twitter for ad spend.
Although there is limited detail in Twitter’s announcement of Twitter Surveys, there seems to be a few areas of potential. Firstly, the native survey format can help to monitor the impact of a brand’s overall activity on Twitter, although we need to make sure we avoid making the same correlation or causality mistakes as in some early attempts to analyse Facebook marketing.
If the format is able to build control/exposed samples based on Promoted Tweet exposure, then we may also be able to gauge the impact of individual promoted ad products. For instance, we would certainly be interested in looking at the relative impact of promoting a celebrity endorser’s tweet about a brand versus what people’s usual social graph are saying about it.
Outside of the link to Twitter’s other ad products, Twitter Surveys also offer a fast scalable alternative to emailing surveys to a list of ‘people who like filling in email surveys’. Although we have instant access though monitoring tools to the global Twitter conversation, I feel that the initial attraction of Surveys is in proactive targeted research opportunities – although again, this will depend on the format of the survey allowing enough data to be gathered, and on the propensity of Twitter users to actually complete the things.
The jury’s still out on whether Promoted Tweets/Trends are a successful marketing tool, and so the premise of Twitter Surveys to measure the impact clearly makes sense. The partnership with Nielsen gives the proposition credibility and the set-up of an independent Twitter account (@TwitterSurveys) to send out surveys rather than the brand itself means impartiality in collecting the data.
The survey platform has been designed to be as seamless as possible for the end user, with as few clicks required as possible. However, I’m not convinced it will provide the answers that campaign developers need. Time is of the essence on Twitter. According to the Wall Street Journal, the average time spent on Twitter per visitor in January 2012 was 21 minutes over the course of the month, compared with 405 on Facebook. Assuming a survey would take 2-3 minutes, that’s 10% of their time during that month, a big ask for brands and for Twitter, especially if surveys aren’t incentivised.
Users are already exposed to plenty of Promoted Tweets/Trends, will they just start zoning out these ‘spam’ tweets, just as they might do ‘spam’ emails? I will follow @TwitterSurveys with interest.
It was inevitable that social media giant Twitter would explore such an undertaking – and it would have been foolish for them not to. It is a great venture that will open up research to a new demographic of people who previously wouldn’t have taken part in surveys.
However from a purist perspective, I think it’s important to ensure that the research is only associated with tweet campaigns. If clients see this as a cheaper means to reach and gain insight from consumers, I personally feel the resulting research would lack real validity and standardisation.
Whilst this service is new, it’s hard to say exactly how this will pan out. I think Twitter may still have some work to do. Although the speed and efficiency of Twitter is great and the idea is very compelling, it seems limited to a somewhat homogeneous audience. Individuals that are on Twitter and engage with a company’s brand(s) via Twitter are often already advocates of the brand and its products thus only providing the point of view of the loyal customer.
Having this view of the base is good, but does not help a company expand into new segments and/or uncover possible issues. You can have an aspect of ‘group think’ here reflecting each others thoughts. While this is very friendly and full of consensus, it is not the whole picture and could be very misleading and biased.
In the UK, only a small proportion of the population uses Twitter frequently; a recent Vision Critical study showed that just 15% of the population had used Twitter in the last week. Twitter’s audience is so niche, clearly this is going to skew survey results.
Unless you want to survey Twitter-based ad campaigns, young urban professionals or high-profile stories such as issues with the iPhone-5 release, studies should only be used in conjunction with other types of research. We’d say it’s an interesting additional channel to gather views.
Twitter’s entrance into survey research, following Google, helps solidify the argument that we are entering the golden age of online market research, especially as the large majority of populations of Western countries are now online in some capacity, including mobile. Twitter’s focus on campaign measurement is more nuanced than Google’s approach, and if Twitter puts sophisticated respondent selection methods in place, the offering could be compelling, linking brand exposure online – both tweet campaigns and online ad exposure – with respondent-based brand recognition and recall – something only possible using survey research.
Twitter surveys also allows for a structured data stream to be part of the analysis, where up to now reviewing tweet campaigns required the complexities of imputing measurement from unstructured text. We mustn’t forget that this foray is a pilot, and as such, while we anticipate enhancements to their offering, Twitter will endeavour to help brands measure and understand the value of tweet campaigns, and will continue to work on its proposition.