Signals in the digital noise
Social media and search data can be informative for brands, but only if they can find meaning in the billions of data points, writes Millward Brown’s Jane Ostler
If you’ve ever monitored a major brand online, you’ll know what it’s like to be dazed and confused.
Thousands of tweets, updates and posts with opinions, customer service issues and brand related imagery will stream through your feed, making analysis impossible.
The volume of data is too much for most humans to process. Technology and analytics tools and techniques are critical to keeping marketers’ fingers on the real-time pulse of consumer behaviour.
The challenge is to identify what is meaningful in the digital noise. In our search for ever greater consumer understanding, it would be foolish to ignore the rich big data potential of search and social.
Making this work, however, means being able to identify the right signals at the right time and being able to attach meaning to them.
The truth is that not every storm in a Twitter tea-cup will have an impact. One-off events such as news stories or PR activity that resonates with consumers online need to be separated from the spikes driven by digital brand or social media campaigns.
Looking at gross signals from search and social will mean that brands are constantly yoyoing around and may over-react to short-term fluctuations, which is not a stable route to long-term and sustainable growth in brand value.
Splitting out the long-term from the short-term trends requires advanced analytics, because big data is noisy, vast and complex. Passively collected and only semi-structured, social and search data provides a signal that is a much less ‘clean’ than traditional models.
Dynamic Linear Models enable marketers and media agencies to deconstruct the raw signals into their component parts. That can reveal how much of the fluctuation of daily or weekly noise around brands is due to long-term factors, and how much is due to short-term factors such as news, events or specific advertising campaigns.
[advert position=”left”]
Long-term strategic brand planning indicators
Long-term trends in search volumes reflect changes in brand salience over time, and levels of active interest in the brand. In social, long-term trends reflect changes in the talkability of the brand.
The reason why these differences matter is because people will search for brands with a high salience; and they will only talk about brands that are worthy of comment.
Identifying long-term signals help marketers and agencies to understand the position of brands within their category as well as the performance of individual brands over time: those which underperform, and those on the rise.
Short-term indicators of campaign impact
The short-term search signals typically reflect how well a brand’s current advertising is driving interest.
Changes in social media reflect how successful a campaign is at driving conversation about the brand, and messages which focus on news, product launches or offers, are more effective at driving search, as consumers seek to find out more.
Both types of short-term signal have been slow to respond to changes in media spend (both positive and negative), but the size of the movement in each can reflect different things about a campaign.
Where a campaign is focused on launching or relaunching a product and is backed by a high level of spend, it can boost search volumes by as much as 9%. Broader campaigns aimed at general brand enhancement with little new news, even when backed by the same investment, will increase search by just 1%.
By contrast, social buzz tends to be more reflective of the creative quality of the message. Comparing the magnitude of uplift in conversation for a given level of spend with Millward Brown’s in-market AI measurements shows much bigger increases for ads with higher cut-through.
Such measurements give marketers an early in-market read on campaign performance, allowing them to make early decisions, optimising or switching out creative that may be underperforming.
From ‘what’ to ‘why’
And so what? It’s not enough just to collect data. Marketers and media agencies need to understand how the signals are moving, but also why – and what is driving them.
The goal is to turn big data into metrics that brands can use to make faster decisions on campaigns and media spend, while also slimming down other surveys so that they are only tracking the metrics they need to, as often as they need to.
Taking this approach helps brands make sense of the mass of signals coming at them from search and social. They can make sense of those consumer moments meaningful, but could also free-up research budgets for more deep dive adhoc and strategic research.
Jane Ostler is sector managing director, media & digital at Millward Brown
Follow us on Twitter: @MediatelNews