So, how exactly do you put a value on social media?
by James Lawson
3 Oct 2012: Understanding a company’s social profile via sentiment analysis and creating a buzz for new products and services remain the two big marketing applications in social media. Techniques for both are moving on quickly, but tracking response and estimating ROI are still primitive. How can you put a value on social media?
How do you feel about us?
The most obvious use of social media is in market research. Sentiment analysis lets you eavesdrop on your customers, and so get a headstart on reacting to new problems or opportunities. At least, that’s the idea.
“Market research is quite an artificial environment,” says James Lovell, Smarter Commerce Solution Consultant at IBM. “In social media, millions of people are feeding into the analysis, giving their honest view. It’s real-time customer research.”
Monitoring tools and services are proliferating, with over 250 of them at the last count. Some examples include Brandwatch, Radian6, SM2, SiteCatalyst and IBM-Cognos Consumer Insight. They work by crawling over the web gathering data and passing it back to their servers, in a similar way to Google’s search spidering.
The giant lumps of data collected this way must then be prepared by stripping out extraneous data. Analysis is hugely complex, involving language detection, keyword matching (does your brand or product feature?) and classifying the tone of comments (sentiment analysis).
IBM client Henkel employs social media monitoring to understand what its buyers are saying about their products, how they compare them to the competition, to measure their share of voice and to glean any new directions for their R&D. One report picked up concerns about the use of silicone in hair care products.
“Henkel was able to address their concerns in its marketing material and they tailed off,” says Lovell. “Prior to the launch of a new product, they were able to identify important weaknesses in a competitive product and so could emphasise their superiority in that feature in their marketing material.”
In another example, Gatorade noticed a lot of interest in its advertising music and was able to offer it as a free download the same day. At Wimbledon, IBM analysed almost three million Tweets, tracking the growing buzz around Andy Murray as he fought his way to the final – and the largest number of positive tweets.
Essentially passive, sentiment analysis is tough to equate with value. At Henkel, it converts share of voice into income. “Through other studies, they know share of voice drives additional revenue,” says Lovell.
The challenge is to bring social media into line with other channel audiences. How representative are those you are tracking on Twitter of the audience you advertise to on TV or send emails to?
“There’s a huge gap,” says Howard Barber, Insight Partner at Tangible, which offers its Soapbox tracking service based on the SDL SM2 tool. “Social media is still in its own bubble. It’s great for reputation and crisis management, and for tactical spotting of potential crises, but is there a bottom line for the commercial benefit? In short, no.”
According to Barber, this lack of demonstrable ROI makes some clients understandably reluctant to embrace social media. “We’re working on bringing more insight and analytic rigour into it, but it’s still theoretical at the moment,” he says.
The aim of outbound social media marketing is akin to TV or poster campaigns: to spread some kind of message and raise awareness of a brand or product. To help initiate this sharing and forwarding, tools are appearing to target those most likely to do so – the influencers.
Through scoring by social influence, Indicia’s Inspire segmentation helps pinpoint these human broadcast outlets. Indicia defines the score as a combination of a person’s likelihood to share and the impact of that sharing, helping to differentiate real influencers from spammers.
“As one of the key drivers for the Tourism Ireland MyIrish140 and Road Trip campaigns, we selected Inspire segments from their visitor and enquirer base that were most likely toadvocate using Facebook and Twitter,” says Julie ááá
ááá Atherton, Planning Director at Indicia. “These segments were vital in starting the initial momentum.”
The data to build the segmentation is gathered on a bespoke basis for each client project using a variety of tracking methods. A representative sample of individual behaviour like forum posting, email opening and forwarding or twitter posts and retweets is analysed and then, in conjunction with other sources like market research, inferred across the whole base. The company has tested the segmentation’s validity by teleresearching a random sample of those in the strongest Twitter segment.
“We found a very strong initiation of Twitter traffic by them,” says Atherton, “You can pick celebs and other key influencers, but this method lets you find more normal advocates.”
It’s also possible to attribute sales value to individual influencers (the value of their individual network). In one retail engagement, Indicia could spot those who sign up for offers and don’t buy, but instead spread the word to others who do.
So you can plan an audience, but how do you track a campaign? A combination of Facebook Analytics and Facebook Connect can report actions like linking to a company website, but Facebook itself is opaque to external tools.
“The problem with Facebook is that everything is hidden in there and they own the data,” says Russell Marsh, Group Strategy Director at Rapp.
Tracking clickthroughs from email campaigns that encourage recipients to “share to social” is one way to estimate social media’s contribution, and Rapp’s award-winning work for Cancer Research’s Race For Life is an outstanding example. The campaign kicked off with emails to Cancer Research’s own database. If recipients shared content on Facebook and their friends clicked on the embedded URL, it took them to the Race for Life website.
The use of a dynamic URL meant Rapp could link back to the original email as the (anonymous) friends arrived at the Race for Life website. Then Rapp could set a cookie and track them around the website until they left or registered.
“We know the exact value of donations plus the influence of each person,” says Marsh. “X might give £20 while Y only gives £10 but influences another 30 people to donate.”
This viral multiplier effect is precisely why targeting influencers is so desirable. To support this aim, new networks and metrics are appearing. Scores from PeerIndex and Klout are based on an individual’s number of posts, followers and reciprocal engagement and how they drive action by others.
Klout takes into account more than 400 variables on multiple social networks to score the individuals that opt in to its scheme. Companies can pay to promote products to these influencers and can also use the score to estimate the success of their campaigns.
Companies are catching onto the value of these ultra-social individuals. “If you have a Klout score over 42, Cathay Pacific lets you use the business lounge,” Marsh reveals. “Virgin give free flights to high scorers.”
As Twitter is even more impenetrable when it comes to tracking and understanding audiences, accessing these influencer networks has great appeal. Where customer relationship permits, another option is simply to use Twitter as a broadcast channel for promo offers and then get buyers to identify themselves in order to redeem.
Amex’s US Twitter promotions are a good example. Here card members had to register their cards to gain discounts via customised Twitter hashtags, Because Amex’s Smart Offers API handles the link between Amex, cardholder and retailer, savings are loaded directly to buyers’ synced cards and automatically pop up on their card statement. On top of that, both Amex and the merchant get to view every detail of individual spending behaviour.
The final part of the puzzle is: how do you define if a campaign has been a success or failure in social media? The amount of buzz or by real income? At the moment, it’s the former in the shape of the number of likes, fans or followers an initiative gathers, but the jury is still out on whether likes equate to extra sales.
“If the only way I can get access to your Facebook page is to click ‘like” then it doesn’t mean I like it, it means I’ve jumped through your hoop,” says Marsh, noting the ease which it’s possible to buy likes or Twitter fans. “The going rate is about $1. Likes are a dead end.”
However, simply counting the number of likes or fans, though definitely flawed, is better than nothing. At the minimum, it gives the agency something to aim for in spreading awareness.
“There are no universal measures of effectiveness,” says Rob Edwards, Head of Insight at 2020. “At the start, you need to sit down with the client and find out what they want or expect from the work. That might be, ‘get me 10,000’ likes or ‘prove to me that Facebook is a profitable channel’. Unfortunately the second isn’t really do-able yet.”
A recent campaign for Mountain Dew gives a flavour of 2020’s approach. Profiling the customer base showed a youthful market with a liking for graffiti, street art and X Games. A campaign that let them design their own t-shirt on Facebook drew 50,000 Facebook fans in months, with over 290,000 likes on subsequent campaigns.
Activity was tracked using Facebook Analytics and by monitoring sentiment analysis. Subsequent work for the Open University involving a Darwin-themed app not only saw one million downloads, but drew 33,000 prospectus requests from new contact.
“It’s about what they do in real life and translating it to social platforms,” says Edwards. “Social is about engagement rather than just making offers, like above-the-line advertising but with direct engagement with the brand. In most cases, likes is the measure.”
The closest most have got to attributing real income is to equate tweets or a set number of likes with what it would cost for the same volume of advertising impressions. However, some recent analysis (reported by econsultancy) indicates they have definite value – at least for some companies.
Bulmer reckons its fans are worth around £200 more per annum than non-fan customers, while ticketing company Eventbrite has worked out that a Facebook share is worth £2.25 compared to £1.80 on Twitter. Hitwise thinks that each new fan acquired on a Facebook page drives 20 additional website visits.
Gaining a customer’s name and address is one sure sign of success, and so soliciting individual sign-up is often the goal. This is the model that Neolane supports with its Social Marketing module.
Fans on a Facebook app are asked to share their details. Once permission is given, the contacts are piped straight into the database. From there, personalised content can be pushed to the app for that user, or posted direct on their Facebook wall. With access to a fan’s other contacts, conventional channels start to open up.
“The challenge is to convert fans to real addressable contacts,” says Martin Smith, Head of Marketing for UK and Ireland at Neolane. “You need to give them something of value to hand over that information. Then you can start to target them using the classic marketing mix.”
Bottom line benefits
It’s still early days for social media, and real returns on investment are hard to discern amongst the noise and hype. However there are very encouraging signs in sectors like FMCG, as figures from Play.com show.
Customers engaging with one or more of Play.com’s Facebook campaigns spent 24% more, with those referred through Facebook spending 30% above average in their first year. Play.com grew its fans from 75,000 to over 350,000, and reported an 80% year-on-year increase in sales through Facebook – worth a cool £2m.
Profiling the Industry: Part 3
20 May 2013: In the third of our special series of articles profiling the database marketing industry, we catch up with Hopewiser, Alchemetrics, Callcredit and greenstone data solutions.
Big data: big opportunity, or big headache?
5 Apr 2013: It is predicted that, over the next year, the human race will produce 4 Zetabytes of new data, which represents about 1 quintillion new objects. Are marketers ready for this deluge and is it an opportunity or a threat, asks the IDM?
B2B email data providers respond to the challenge
12 Mar 2013: B2B email data providers have gotten increasingly sophisticated over the last decade when it coems to building compliant, legal and responsive lists, as James Lawson discovers.
Game, set and match
12 Feb 2013: Efficient name and address data matching is a fundamental of much database and insight-driven marketing activity, even if it doesn’t inspire much excitement, finds James Lawson.
Should addressing data be free?
4 Feb 2013: With a call for an Open National Address Dataset, the Open Data User Group has stirred up a heated debate on whether addressing data should be made freely available, finds James Lawson.
Sorting out the best mailing options
19 Dec 2012: The UK postal market is now the most de-regulated in the EU bringing both positives and negatives – James Lawson speaks to expert MSPs and vendors to get the low down on sortation.
Trigger happy: market 'never better for trigger data'
18 Dec 2012: When it comes to effective direct marketing, timing is everything because pulling the trigger at precisely the right moment is invariably the difference between a sale and a fail, says James Lawson.
Embracing the challenges of going global
14 Aug 2012: The attraction of leveraging the power of the internet to target overseas markets has spawned a surge in international activity from the UK’s many address management providers, but the challenges of going global can be very significant indeed.
SCV is key: so what's stopping you?
13 Aug 2012: The overwhelming majority of marketers regularly accept that building a single customer view strategy is utterly vital in today’s complex, multichannel world – so why is it that so few of them are actually doing anything about it, asks James Lawson?
Cleaning up at both ends of the market
26 Jul 2012: Aimed ostensibly at SMEs, online self-service data cleansing sites are increasingly pitching themselves at much bigger customers looking for convenient, quick turnaround options – so can that circle be squared, asked James Lawson?
Leveraging web data remains a challenge
18 Jun 2012: The torrent of data made available by the web is hugely tantalising, yet harnessing it for use in multichannel marketing remains a huge challenge, finds James Lawson.
Is big data really such a big deal?
14 Jun 2012: The buzz around ‘big data’ has reach epic proportions, but do marketers really need to get so excited and concerned about it in the UK, challenges James Lawson?





