John Griffiths, Planning Above and Beyond, reports on WOM UK's 'How To' Word of Mouth Marketing Workshop. Among the topics under discussion were how to identify "influencers", using brand agents, establishing a suitable budget and how to measure the payback from word of mouth. But John opens his report with a mini-experiment all of his own ...

Listen to a podcast where speakers from the Word of Mouth Marketing Workshop discuss the presentation immediately prior to their own.

Listen to a second podcast where the speakers from the event assess the talk which came after their own.

Look who’s talking. And laughing

This is the first event where WOM UK (which is still less than a year old) has set out the practicalities of developing word of mouth strategies. One of the most interesting aspects to a day like this was going to be learning how sound the theoretical basis for word of mouth marketing was and how readily it could be applied.

For many years now, the business press has carried stories about the growing significance of word of mouth in helping people to make decisions. But that in itself doesn’t tell you how to do it effectively, and particularly whether companies can encourage people to recommend on their behalf. Last week, I saw a statistic that two thirds of all Facebook users are happy to have brands listed among their friends. To which I can only answer “Bunkum”.

Steve Barton, President of the WOM Association and the chair for the day, began with an overview of the word of mouth – dramatic WOM campaign results such as the 1156% sales uplifts on pet food on Tesco’s Clubcard database, the study that showed that one piece of advocacy could generate 62 conversations, and the fact WOM is worth $1.27 billion dollars in the US and £367 million in the UK.

Some of this reminded me of IAB gatherings of the faithful back in 2002–03 quoting billion dollar stats for the prospects for digital advertising. And some of the numbers raised eyebrows then. But who’s laughing now?

Identifying the influencers

Then Fiona Blades took the stand for the first paper, which was about identifying “influencers”. I had already seen her presentation for Boom Chicka WaWa and the work Mesh had done for Lynx, from which this commercial comes. 

But this time round, Fiona included new findings, and started a theme which was to be referenced throughout the day by asking: Who are the influencers? The answer seemed to be everybody, but some are more influential than others. As such, choosing whether to amplify through ordinary people or to go and look for the “mavens”, “connectors” or “salesmen” who will do the job quicker and more credibly is one of the fundamental questions.

Fiona came down on the side of finding and measuring the impact of the many, and had the results to demonstrate it. Its is complicated, though, because Mesh has now added another seven roles to Malcolm Gladwell’s original typology, which all overlap but which are interesting because they represent the modes the average person works through in being “credible”.

Using existing word of mouth as it happens – rather than trying to work via a small number of expert ‘pushers’ – points the way to measuring WOM in a far broader territory than the hothouse of a handful of picked mavens who would always have more to talk about than there was time to speak of.

The importance of self-identity

It was then on to Bryn Bazzard, of Quocom, for the second presentation of the day. He addressed the targeting issue from a data perspective, and was critical of the TGI questions, which have been on the BMRB survey from time immemorial, and really aren’t helpful for the kinds of segmentation which WOM marketers require.

He also helpfully covered some potential areas of confusion, one being that key influencers are not necessarily heavy users or experts in the particular category of interest, but do like to talk about it. He then identified the three main methods of profiling, developing a scoring model (which had the benefit that it could be tuned based on influencers subsequent behaviour) and whether or not to have this group “self-identify”.

He concluded with a three-way model where the influencers’ self-identified credentials could be correlated with qualitative data about their interests and combined again with quantitative profiling. The challenge is whether all three can be done economically; this sounds rather like a Rolls Royce solution requiring checks from three different angles.

Applying social glue

After coffee, the next session was about identifying the appropriate content. This was led by Andy Crysell of Ramp. This was a researcher’s perspective on content, and therefore naturally considered how content could be researched.

The basic insight in Crysell’s paper was that the role of content was as a sort of social glue: this was how content is judged by those who use it and receive it. Marketers, however, often have a very different perspective.

This gave rise to a number of examples of research projects where there had been a significant shift away from the conventional focus on ‘content’. He gave the interesting example of the L'Oréal swatch for choosing colours for hair colouring, which requires those using it to ask for the help and opinions of their friends. The conversation comes about because of, and through, this interaction, not primarily because the topic which matters to the marketer is similarly compelling to the customer.

Turning agents into mechanics

The last session before lunch featured Ivan Palmer of Wildfire and Dave Balter of BzzAgent. Both of these companies are using word of mouth as a kind of media channel – you buy agents by the hundred (or thousand) and pay the agency for the promotional of your message.

It represented a certain divergence of philosophy, since Wildfire recruits people as influencers for specific campaigns in which they are interested, while BzzAgent has a substantial panel of agents (400,000 in the US alone) who are selected and deployed for different tasks but may be on the books for years.

Ivan’s presentation, which was one of the most substantial of the day, covered how to build what he called an accelerator – a “mechanic” which would enable influencers to effectively communicate a branded message.

This involved five criteria – that it built up the status of the influencer, that it involved a discovery of some kind of new information, that the conveying of it was in the interests of the person receiving it, that the occasion was relevant and indeed that the content improved the occasion, and, lastly, that the platform made conveying the information natural and easy.

Unless all five criteria are satisfied, success could be elusive. And, actually, it is hard to build an accelerator which naturally does all five. As such, and using the Pringles Rice examples, Ivan showed how a number of different activities ticked all five boxes, though none always delivered on all five. This raises a question for evaluation: what proportion of the target would be exposed to all of the materials?

David Balter’s presentation was a straight introduction to the business of using agents. He contrasted buzz marketing with shill marketing. The latter uses scripts and paid sales agents who endorse the product; in the BzzAgent model, the agents are not paid – they volunteer to participate and are given briefing packs, but they aren’t given scripts and don’t even have to be complimentary about the product: they are even free to point out weaknesses.

This was also a very full presentation. Probably the most interesting part was the admission that since beginning the company, it had moved its position from trying to identify and motivate key influencers who would be more influential for everyone, to motivated people who were likely to be influential from time to time in very particular context.

This brought Balter much closer to Ivan Palmer’s position. All in all, this would have been a very helpful session for any clients wanting to buy into word of mouth as a comms channel.

Measuring WOM

After lunch, it was time for measurement – not the most sparking topic for digestion! Mark Rogers, of Market Sentinel, started to unpack the key metrics for evaluating campaigns. Inevitably this came back to determining the Net Promoter score, and trying to correlate the importance of the volume of traffic which word of mouth had driven versus the authority of the influencers.

Every campaign will deliver a mix of the two, but determining and, in particular, predicting the balance between them is very difficult. Mark said that anything from the government has high authority scores, and likely to be believed (I wonder if that it always true?)

Likewise, he argued that bloggers generate lots of traffic, but don’t have much authority (though that may well depend on the individual blogger). Mark then produced some case studies which showed how by researching sources to determine their trustworthiness: Avis, for example, had increased its trust scores by getting itself talked about by those with whom its values were strongly associated.

Budget setting

Richard Moss, of Weber Shandwick, followed with the unenviable task of explaining how to set a word of mouth budget. This was a basic and easy to follow presentation, and showed how one might go about such a task. Its weakness appeared to be the dependence on respondent reporting: How do we know how many average recommendations an influencer makes? What proportion of people trial and how many times are they communicated with before they tip?

It was a little reminiscent of reading early Drayton Bird emanations where you kept wanting to ask: Do they always do this? How many studies was this drawn from, and what were the base sizes?

But as a rule of thumb for working out budgets it was helpful (I just wouldn’t stake my life on the cost per impression!). What it did show was if the figures are reasonably accurate, the investment in Word of Mouth marketing is so affordable – in the region of a typical qualitative research study – that the only reason why a client shouldn’t try it is a) ignorance and b) lack of imagination. This is not an expensive area to trial relative to other marketing costs.

Putting theory into practice

And so on to the last session of the day, and a workshop led by Fraser Chisholm of Royal Mail. This was a practical session, where the delegates were split into three syndicates and worked on developing a WOM strategy (including identifying influencers, the appropriate content and an appropriate accelerator package, and so on).

After all that sitting and listening, it was remarkably difficult to get into active mode and start to put what we had heard into practice, particularly when under time constraints. It was also very easy to argue over the definition of the audience and influencers – both of whom are of course critical to the success of the programme.

Looking back, and the way forward

And so to the end of the day. In summary, I was agreeably surprised: there was real substance here. The challenge is that, like most emerging disciplines, word of mouth draws heavily on other established marketing territories like direct marketing and digital to explain how it works, when it is actually quite different from most of them.

What we have here is a horseless carriage which is going to turn into a car within a relatively short period of time; then people will forget what it was like when the basic theories and working practices were being laid down.

The word of mouth market is here, and it will be accountable in theory and practice. There are, of course, some dangers of burnout if it is treated as just another channel which can be switched on and off. I can’t believe that influencers, even if unpaid, will be content to do this forever, and those who are may be the kind who most of us would rather avoid – rather like people who can’t stop going to focus groups.

One last consideration as WOM adds one more redefinition to the maelstrom which continues to engulf marketing. Why can’t WOM projects be used as alternatives to test markets as a kind of research/test marketing hybrid? It isn’t research, because it can be shown to pay back.

However, for gathering market intelligence of the most valuable kind, the budgets involved are similar to small scale research budgets and arguably it delivers market insight. Why couldn’t word of mouth be an alternative to research?

There is still an issue with clients committing any budget at all in this area. No one would admit this on the day: they have their sights set on the big time – targeting sales directors and procurement staff looking to amortise media costs. The danger with this approach is that it throws away a lot of the benefits of word of mouth: genuine engagement benefits for both the influencer and those influenced.

If WOM goes mass, it will also have a burn rate. I just wonder if an acceptable way for WOM to grow would be starting small and being willing to grow slowly, embracing small scale and paying its way from day one.

A report on WOM UK's "How to" WOM workshop is coming early next week.