Follow EmergingSpaces on Twitter

Insight, news and opinions from Starcom MediaVest Group London.

Subscribe by RSS or by Email,
and follow us on Twitter.

Actions speak louder than Clicks

Over the last few months a Starcom MediaVest Group taskforce of our social, digital, data, search and research experts has been on a mission to create a simple yet consistent and scalable set of evaluation metrics for social media.

Our starting point is that the key to understanding the true effects of social is to not just look at the numbers participating, but the level of their involvement. Also, we believe that if we want widespread adoption of our approach we must utilise existing data feeds from the social media publishers and not rely on expensive third party sources or black box models.

Clearly, the more involving the social media experience within a particular channel the better the brand experience. Unfortunately, the available data does not provide any evidence of a link between involvement and actual brand actions that move the consumer closer to purchase.

To address this issue SMG has just conducted a major research study to quantify the value of social media experiences from exposure through to purchase funnel actions. In short we wanted to discover the propensity of someone who perhaps may write a positive comment on a brand Facebook page to then go on to do something of tangible value to a brand, such as consider it, talk about it to others or visit the brand website.

In all we observed a representative sample of 6000 regular Facebook, YouTube and Twitter users and asked them to spend several minutes interacting with content on brand Facebook and YouTube pages in product categories that they had already registered an interest in. Once they had finished, we asked questions about what further engagement they might make with that brand on that social media channel, as well as any possible brand actions they might take in the future. The combination of these two sets of information gives us what we call the Social Media Behaviour Index (SMBI).

Picture1.jpg

In order the make the study as robust as possible we collected information on 6 product categories and 29 individual brands.

The results are very revealing. The key headline is that people who do more than just click on a branded social media site (i.e. play a game, post a comment, would follow on Twitter etc) are twice as likely than those who just visited the site and did nothing else, to do a purchase funnel action ranging from enquiry to preference. As you see in the table below the figures for Facebook were slightly higher than YouTube.

Picture2.jpg

The SMBI highlights the value of social media actions by indicating future brand actions, and identifying the characteristics of people who do them. Put simply, the higher the SMBI, the better social media behaviours are at expressing the value of Facebook, YouTube etc to a brand. Crucially for our work, we noticed that across the different brands measured the likelihood of business specific brand actions following social media behaviours varies from category to category.

Picture3.jpg

Finally, the research dug deep into how specific social media behaviours translated into a greater propensity to future brand actions. We found out that on Facebook tweeting content or following on Twitter correlates with the strongest brand actions (i.e. 9 out of 10 will follow up with a subsequent funnel action) although watching a video, posting positive comments or passing to others still generate high conversions to funnel behaviours except brand preference.

Picture4.jpg

We believe there are a number of important conclusions with corresponding implications from this research. Here are just a few.

• Social media involvement leads to brand behaviours that do much more than improve awareness or drive interest

• The deeper the interaction with social media the greater likelihood of moving the consumer from enquiry to preference

• The most common brand action after a social experience is to leave a brand’s page to go to its website to enquiry so it is very important to know which categories and brands need to make Facebook the final destination

• Tweeting and following via Twitter appear to be behaviours that correlate most strongly with future brand actions so make following a brand via Twitter easy.

• Video, games and competitions are more useful for creating ongoing relationships and loyalty than directly influencing purchase funnel actions

• Keep social media content fresh as interactions are constantly changing

For the SMG team building a new set of social media evaluation metrics this research has done more than just substantiate our belief that activity in the social space is about creating “Likes“ AND subsequent actions. It has also provided the data points to optimise social media objectives with brand action outcomes which is essential information for our proprietary Experience Involvement models.

2 comments

Mark wrote 1 year 33 weeks ago

Hey guys,

Firstly, well done on some great work. I just had a few questions:

How did you select your sample (to ensure it accurately reflected the true population)?

How did you decide on 6000 respondents?

Did you ask respondents the questions or did they read from a questionnaire?

Did you use any statistical test to check the significance of your results?

Steve Smith's picture
Steve Smith wrote 1 year 32 weeks ago

Hi Mark

Thanks for your interest in this work. I undertook this study, so here goes:

The whole sample was selected in part to be representative of people who are likely to visit a brand's Facebook page and/or YouTube page. In the first instance, then, each respondent already needed to be a regular Facebook and YouTube user. Respondents for each brand were selected partly according to their interest in the product category relating to the content of that page e.g. visitors to the Heineken brand Facebook page needed to be a regular lager drinker (among other factors related to that brand). The idea was to make sure each respondent already had an existing likelihood to visit that brand page. We are aware of some of the issues related to this method (and where possible have included these in our analyses and as caveats), and are looking at alternative approaches.

6,000 respondents relates to the number of brands we had surveyed. As we continue to include brands, then so the sample size increases as does the overall accuracy of the metrics we deliver.

The questionnaire was an online questionnaire.

Regarding statistical significance, yes we have, and the statistics we have used include Chi Square, and correlation and regression. We are still analysing results, and we are looking to include some loglinear regression and factor analyses.

Please feel free to get in touch (steve.smithsmvgroup.co.uk).

Steve

Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.

Add your comment

The content of this field is kept private and will not be shown publicly.

More information about formatting options