Department store John Lewis have announced that whilst their rivals may be suffering, they are growing market share. The Nationwide building society has seen a surge of deposits recently, partly the effect of the Northern Rock collapse.

Both businesses are growing market share during a slowdown. What is it that links them? In news links Nationwide talk about the quality of their assets, John Lewis MD Andy Street is not quoted as providing a rationale.

Our research points out some common features. What both businesses have in common in terms of consumer generated conversations is that:

a) positive commentary on them tends to contain specific customer recommendations and endorsements. A customer who is complaining about his ISP takes time to say something postive about Nationwide, an entire thread on MoneySavingExpert is entitled John Lewis are bloody marvelous and backs it up with facts;

b) negative commentary involves isolated problems: someone complains about a silent call from Nationwide’s call centre; a thread that starts John Lewis sucks big time, turns into a plug for their customer service as - just as several posters predict - John Lewis deal successfully with a horrible customer service issue.

This is by no means a common feature. A PR client came to us a few days ago on behalf of a business whose online commentary was positively sulphurous. There were no positive comments whatsoever, and the negative comments included threats of legal action. The company apparently thought that it had a “reputation” issue. Our suggestion was that it had a product issue. This is an ostensibly healthy company, but I would fear for it during a recession.

The common threads linking Nationwide and John Lewis is that they seem to provide great customer service and great customer service drives positive word of mouth. Both companies Net Promoter Indices are comfortably ahead of their sector average. And they are demonstrably growing market share in a chilling market.

When we helped Avis launch their We Try Harder blog - it was a joint venture between customer service and marketing. The point that Xavier Vallée and his colleagues at Avis understood is that customer service issues - correctly handled - are the key to having a great reputation. No one is perfect, but if your service is responsive and prompt, you are forgiven and endorsed. Avis, like Nationwide and John Lewis, are growing their market share.

We published this press release today:

AVIS UK WINS INNOVATION AWARD FOR SOCIAL MEDIA PROJECT

Avis UK has won the SOCAP award for innovation in Customer Service at the National Customer Service awards dinner held in London on 18th September 2007.

Avis UK’s work involved monitoring and benchmarking consumer generated content and making changes to its product and customer service practices in response, culminating in launching the We Try Harder blog. These changes have resulted in big increases in the approval rating for Avis (the Avis Net Approval Index has increased over 200% and remains well ahead of the competition) and key product innovations.

Market Sentinel provides the technology for CGC research and guidance on blog editorial.

Market Sentinel CEO Mark Rogers said yesterday: “The phenomenal growth of social media over the last few years has left many companies scratching their heads. Many congratulations are due to Eibhlin Payne and Xavier Vallée of Avis UK for demonstrating to others how to use statistical methods to understand social media and develop insights enabling them to forge real links with customers. We are very proud to have contributed to Avis UK’s success in creating a genuine partnership between marketing and customer service. Many thanks are due to Sheila Sang and Caroline Harris who worked so hard in helping Avis launch the We Try Harder blog, to Mathew Vattolil whose analysis drove decision-making and to our partners at Web Liquid, Matt Cronin and David Shiell who had the vision to bring the project to life, to Xavier Vallée and Eibhlin Payne themselves for commissioning the work and to Rob White for his tireless efforts in bringing it to reality”.

We shared the Dell case study showing the apparent association between customer recommendation and stock price with Peter Hutton of Brand Energy Research writes:

Thanks for sending me this.

Yes, I am familiar with Reichheld’s NPI. Curiously enough, I developed my own advocacy scale, with a net advocacy score, 12 years ago, before Reichheld came to prominence with the book ‘The Loyalty Effect’ published in 1996, I believe.

The NPI has been much criticised not least recently by a chap called Tim Keiningham who looked at Reichheld’s own data and found that the conclusions he asserted in his book just did not follow from the data and analysis that he had used!

This did not surprise me too much since, when I was at MORI, I tried unsuccessfully to find positive correlations between net advocacy and growth (of profit, sales or whatever) and concluded it was for two main reasons:

1. It is rare, and generally impossible, to find survey data and financial data that also have exactly common fields and in sufficient quantity– e.g. time period and business unit – to run robust correlation analyses. Customer satisfaction surveys are often conducted in discrete time periods that do not correspond to the business accounting periods. Moreover, they are normally undertaken for business sub units or products/services whose financial performance is not necessarily separable from other areas of activity in the business’s accounts and/or where spending in one period is not necessarily designed to have an impact until some time later (e.g. R&D, recruitment, marketing etc).

2. What drives value creation in a business is very complex. In the long term, of course, more loyal customers will enable you to grow faster; but so too will a lot of other things – properly structured financing, favourable market conditions, weak competition, well motivated staff with the right skills, clear business strategy, effective advertising/marketing, sound leadership and management and so on. All are necessary but none are sufficient on their own to achieve outstanding performance. In that light it is no surprise that one factor – customer loyalty as reflected in the NPI – has no correlation with growth.

However, that does not detract from the value of the idea underlying the NPI. In my experience knowing that you have x% advocates and y% detractors is very important and should be a key focus for top management. But the most valuable questions, as you imply in your piece, are those that reveal what is driving your scores up or down. They are the only ones you can actually act on and are therefore arguably the Ultimate Questions.

This is what I wrote to him in response:

Thank you for this very thoughtful response. We have a lively debate on the NPI with Justin Kirby of “Connected Marketing” fame. He makes the same point as Tim Keiningham, broadly.

Fred Reichheld may have over-stated the case when he entitled that book “The ONE number you need to grow”! Clearly you need to have other things in place to execute. A company can promote warm feelings but it takes marketing professionalism to turn that into a world-beating brand.

My hunch is that online monitoring solves some (but not all) of the sample bias issues associated with using NPI off-line and that it adds an interesting additional value which is that it provides an index of

A) voluntary, spontaneous endorsement or criticism
B) explanations for that endorsement

which goes very much to the heart of what businesses are looking for when they want to create word of mouth marketing campaigns.

Dell have just announced a profits jump of 46%, topping analysts’ estimates. The stock price rose to the $28 level. This chart shows the stock price’s movements over the last six months. [Data from Marketwatch]

In March 2007 Market Sentinel completed a case study looking at how sentiment was changing relative to Dell’s customer service. It is Frederick Reichheld’s contention that you can predict the future of a company by asking its customers if they would recommend it. We apply this methodology to consumer commentary online, but with a twist. We look for the reasons consumers give for their recommendation or their disparagement of a company. Our study showed an improvement in general sentiment about Dell’s customer service in 5 of the 7 key metrics we used. Customers were still irritated by the simple errors made by customer service in recording names and addresses and by the practice of off-shoring in general. But in relation to general customer service issues, speed of response, technical competence and other miscellaneous issues, sentiment had moved positively from the year before. There was also a highly positive response to Dell’s social media initiatives StudioDell and IdeaStorm.

If you have bought the stock at $22 six months ago on the back of this study, you would have been looking at an appreciation of 27%. That is ahead of the market (the Dow Jones Industrial average is up around 16.6% over this period). Our report lends additional support to Reichheld’s thesis: investors should track sentiment as a leading indicator of company value.

Today we publish a new white paper: Responding to Crisis Using Social Media. It is an update to our white paper Measuring Blogger Influence, which looked at the Dell Hell débacle and measured the role of bloggers in creating the damage to Dell’s reputation for good customer service. Dell has publicised their increased investment in customer services and has launched the social media initiatives Direct2Dell (their blog), StudioDell and IdeaStorm to increase dialogue with its customers.

Has it worked? We surveyed customer commentary from before and after the new initiatives and used our net promoters methodology to find out.

The bad news is that the increased spend on customer services ($150m - on Dell’s figures) has not yet had a strong positive impact on overall sentiment. There has been a slight dropping off in the volume of negative commentary about customer service, but errors seem to be up and opposition to off-shoring has increased. The good news for Dell is that its social media initiatives have offset this, and there are signs that they may be successful in changing opinions about the company. The recent PR wins which saw Dell salvage XP for domestic customers and announce the launch of a Linux Ubuntu desktop (although they occurred after our data sample was taken) have reinforced the impression that using social media is a big customer service plus for Dell.

In Tuesday’s FT Chet Henderson of Unilever Insight (the multinational’s research arm) discussed with Carlos Grande why the company had taken the decision to move 80% of its market research to the web. “Internet response are more honest than those gained by traditional methods,” Henderson observed. The article points out that the market researchers hope to save 10% to 20% of the cost of traditional methods. Grande estimates Unilever’s annual market research budget at €400m (£272m) which represents a big endorsement for the credibility of web-based research methods. The article appears to be talking only about focus-group work. There is a lot to be said for such methods, and companies like Baldo Faieta’s Synthetron have some interesting solutions for establishing focus groups online.

Unilever could also consider making use of the data that consumers volunteer themselves online in blog, messageboards and review sites. The advantage of using consumer-generated content to generate insights as to consumer tastes is that it really does give you an honest view of how the consumer feels - as the opinion has not been elicited by a leading question from a market researcher. The companies who use our reporting are somewhat shocked by how rude consumers are online, but once they have recovered from their immediate defensive recoil, they find the honesty of great value - and of course the same consumers are just as rude about the competition! [hat tip to Flemming for spotting this!]

A client, reviewing some of the sentiment scores on the net approval work we have been doing for them recently asked: “why are these people so negative? We don’t get these scores from our off-line net promoters work.”

The client was identifying a pattern we see quite regularly. Online commentary is more negative than off-line. Why? There is no systematic answer to this question, but if you were to attempt an answer you would divide it into three sections:

The squeaky wheel. People are likely to post on a blog or message board or a consumer review site when they have a reason to. The post can often be prompted by a problem they wish to solve, for example - difficulty with the product - if a car won’t start or a program won’t install. Because of this, and given the likelihood that the problem may persist, complaints about the product and the brand are likely by-products of a web session.

Monologue is more negative than dialogue. Research we have previously published  refers to work done by Delahaye pointing out that blogs are more negative that messageboards. 23% of blog commentary is negative, compared to 11% of message board commentary. The reason? People tend to be more measured, more polite face to face than they are in monologue. They do not make such bold, inflammatory “look at me” negative comments. The reason is that in dialogue a speaker is unsure of the feelings of the interlocutor. If he or she makes an emphatic statement about a product or service, he or she risks spoiling their social relationship with the other speaker (or poster) who may be a big fan of the product in question. This politeness factor may also explain why the results of face to face conversations are less negative than a sample of online opinion might suggest.

The third reason is that, particularly for bloggers, staking out one’s social territory online involves a certain amount of display activity - particularly for men! A vehemently negative comment about a large brand demonstrates a certain kind of alpha male aggression. It is show-off behaviour. Bloggers want to be linked to, and being showily negative, particularly in a witty way, may garner more links than a considered “one the one hand, on the other hand” approach.

Paradoxically, as we explained in our “Measuring blogger influence” report, this negativity amongst bloggers actually reduces their authority on a topic. For the casual reader, strongly negative views are off-putting. A neutral is more likely to be impressed, and influenced by the argument of someone who has weighed up the evidence for and against a brand, citing their evidence, than by someone who says: “Brand X SUCKS!”So … although brands are worried about extreme negative commentary online, it generally has little influence. That is, unless there is a special case.

The special case is, as in the example of Jeff Jarvis and Dell:

a) the issue complained of is real and the complaint justified, and echoed by others;

b) brand makes no direct statement on the topic, leaving the existing authorities no choice but to link solely to the complainant.In that one case, bloggers can be VERY influential.

A lot of people visit this site on the search term “brand audit”. Even with the miracles of Google analytics it isn’t always possible to tell who they are and what they are looking for. If you came here searching on “brand audit” or you are a brand owner interested in “buzz tracking”, here is a short scratch sheet on how you can use companies like ours to analyse the competitive DNA of your brand. Market Sentinel answers the following questions:

Who are the stakeholders in my brand? The partners, customers, regulators who have an opinion about it? Which of them talk about it, with what degrees of authority? What words do they use? How can I address them? Change the climate of opinion?

How do my customers rate my brand? What aspects of it do they particularly like, and which do they particularly dislike? Which characteristics of the brand are key to my customers recommending it to one another? How does my brand compare to other brands?

What is new here is the ability to answer these questions using reliable, tried and trusted mathematical techniques. What is new is the ability to reliably benchmark one brand against another, and to do so repeatedly. Market Sentinel performs this kind of analysis for Avis, Intuit, GSK, Rio Tinto, BP, Hyundai … It can be applied in any industry sector, or to any brand large or small. Our customers are in corporate communications, e-commerce, word of mouth marketing, customer service and market research.

Set up times vary, but you can get expect to get actionable data within two-four weeks.

How do you measure word of mouth? The increasing importance of social networks to brands and advertisers has raised this problem very sharply in the last few weeks. Media owners, pharmaceutical companies, automotive manufacturers all need to know the same thing: how am I doing?

If a brand can establish how it is doing in “word of mouth” in relation to other brands this information can drive decisions about the allocation its marketing or campaign spend.

We would need to agree what a ranking in the word of mouth market means.

In the Market Sentinel methodology there are three possible ways of ranking in word of mouth:

• Buzz (numbers of citations)
  
• Approval (sentiment compared to benchmarked competition)
  
• Authority
  

“Buzz” is chat, pure and simple. Measuring it gives you an indication as to whether something is worth mentioning. Britney Spears has buzz. It ebbs and flows. The weakness of buzz as a measure is that you can be talked about without your product necessarily being purchased, or your value endorsed. Key to consumer brands is the central flaw that not all brands have talk-about-ability. Some products and brands are worthwhile, and do their job well (like car insurance) but they are just kind of boring. That doesn’t mean that they are bad products, or that they aren’t relied on, but it means that using “buzz” to track them is bound to fail.

Approval is better, as it equates to the likelihood of customers to recommend your products. We measure it using the “Net Promoters’ Index” - that is a simple index of how many people promote and how many detract from a product or brand in relation to industry benchmarks.

Authority is best of all, it equates to trust, which means that your marketing messages are more likely to be believed, and it corresponds approximately to Google ranking, since it relies on authorities citing you and linking to you. We measure this using a “Stakeholder Analysis” - an index of all the stakeholders in a topic as to who they view as authoritative.

Here is how we would address the problem of benchmarking a brand in relation to word of mouth:

• Identify the topic in which the brand seeks greater authority
  
• Benchmark its existing authority (conduct a stakeholder analysis)
  
• Identify the key authorities to whom the brand would need to communicate its broader proposition
  
• Profile those authorities in terms of their ego-net (who they link to and by whom they themselves are cited), and in terms of the statistically improbable words they use (i.e. their idiolect or individual language)
  
• Assess the brand’s own “Clarity” (consistency of message on the topic) and “Resonance” (the extent to which the brand’s language is picked up - or ignored - by stakeholders). I will return to these concepts later and examine them in more detail
  
• Then we work with the brand’s communications people to produce high value content designed to appeal to those authorities. This communication material could often be contained in a blog, but could form the kernel of a buzz marketing campaign, or a strategy for offline communications.
  

The last few days have seen us talking to a number of different companies about whether it is possible to measure the success of marketing campaigns using our tools. It is.

The traditional way of measuring marketing success is:

1. Sales of the product. This is the point of the exercise after all. But sales numbers are a crude tool. How and why did these folk reach you? What were they responding to? The original campaign, or word of mouth generated by it? If they are buying it, are the also recommending it?

2. Consumers seeing the campaign. In offline this would include estimates of people seeing the campaign in papers and on TV, in online you could measure how many people reached any of your pages, downloaded your collateral.

3. Numbers of links to your campaign. This is the primary traditional buzz measurement. A campaign needs links from those who are qualified to endorse it.

4. Numbers of citations (how many people mention your campaign). The secondary traditional buzz measurement. If they are talking about you more after the campaign then you know the campaign is working at least on some level.

All these metrics deliver some value, but they don’t really tell you enough about the campaign itself. If you are a mobile operator the retail returns on handset sales tells you what is hot and what is not, but it does so retrospectively. You can’t predict the success of a product on the first day it is sold by a competitor.

The numbers of customers seeing the campaign is a weak indicator of whether a campaign will be successful. The bubble era was strewn with campaigns which were seen by millions but acted on by very few. (Remember the ISP “Breathe”, whose beautiful, but baffling advertisements were hard to avoid?)

The number of links to a campaign and the number of citation of it form the beginning of an important measure. But both these metrics are incomplete in themselves. Firstly not all links have the same value: some links are more equal than others. If a link is from someone who is authoritative in the area of the product, it is worth far more to the campaign, brings more Google influence and hence more prominence in natural search and therefore more sales.

Natural search is the key target for this kind of buzz measurement. Why? When customers go to Google, their eyes disproportionately linger on natural search. Consumers consider natural search authoritative. The image below is courtesy of Eyetools and demonstrates this in practice.

This is authority is inherent to a brand and it is something that Market Sentinel can measure. What can be measured, as the saying goes, can be managed.

Secondly these measures leave out the quality of the citation - is the reaction positive or negative? A citation is not necessarily an endorsement.

We have recently been looking at a clever viral campaign for Sony Bravia on behalf of the agency Tonic.

The campaign’s origin was a commercial shot in San Francisco on 26th July 2005. Thousand of brightly coloured plastic balls were dropped on a street in Telegraph Hill. This spectacle was so astounding that passers by stopped to photograph it. They posted their pictures on Flickr. This photo has been viewed 245974 times. The agency spotted the opportunity for a viral marketing campaign, and created a site using high quality assets from the photoshoot, as well as making the commercial itself available online.

We were not able to give Tonic all the help we would have liked to, because they only came to us in the last couple of weeks, and benchmarking this in retrospect would be time-consuming. However: this is what you would do to measure the impact of a digital campaign like this.

1. Benchmark the “approval rating” of the product. Compare it with two or three direct competitors. How many people recommend and how many disparage the product? Measure this before the campaign, during the campaign and three months later. If you have changed sentiment about the product relative to the competitors, the campaign has had an impact.

2. Measure the influence of the Sony Bravia brand in relation to the sector as a whole (LCD TVs). That is: how many authoritative commentators in the area of LCD TVs give authority to Sony Bravia. To do this you need to use our proprietary software. Again benchmark this against competitors in the sector. Repeat during the campaign and six months after (these effects take longer to worth, and are themselves longer lasting). This is the key area. If the campaign is successful Sony Bravia will rise up the Google natural search returns for “LCD TV”.

3. Undertake a stakeholder analysis of Sony Bravia itself, before and after the campaign. Look at who were the key influencers before and afterwards. Examine the actual words which are used by those influencers. Examine the context of the conversation - what other websites were cited in that context? These subtle and sensitive measurements help you measur how the climate of opinion around a brand has been transformed.

4. Measure search volume for crucial terms. How many people are looking for the campaign, or for the brand in the context of the campaign.

The preliminary work we did looking at this Sony Bravia campaign identified some interesting methods by which the campaign achieved traction as well as throwing up some interesting lessons about what the brand should try to achieve next time. Tonic will be publishing more detail in the next few weeks and when they do we will return to this subject.

We are working with a customer in the automotive sector looking at commentary in both message boards and blogs. It has brought out some interesting characteristics of web users. The volume of commentary in blogs is somewhat lower, and that in message boards somewhat higher than we had anticipated.

I am not aware of any research which compares blog-writers with message board contributors. It seems to us that there are distinct contrasts. Bloggers tend to solitary, opinionated, contrarian, message board contributors are friendly, clubbable and consensus-seeking. Bloggers are cats and message board contributors are dogs. Not very scientific, I know. I am always quoting some Delahaye research which showed that 23% of blog comments are negative and only 11% of message board comments. Previously I had assumed that this was because of the nature of the conversation. The blog is a monologue (with interjections) and the message board is a dialogue, or even a public meeting. Now I am beginning to think the differences go deeper.

We will look for chapter and verse.

New Communications Blogzine has published Mark Rogers’ article looking at how Market Sentinel measures corporate reputations with our “net promoters index”. Here it is:

A year ago when our new company Market Sentinel started distributing live reports on brands drawn from monitoring message boards and blogs, one of our early pitches was to a research head of a large mobile carrier who was sceptical as to the value of it.

“Who are these people?” he asked, reviewing a page of sometimes negative remarks on a new mobile handset. “Isn’t this just chit-chat?”

We realised then that we would need to produce a metric which showed – in as objective a way as possible – how positive and negative commentary played online. We lighted on the work of Bain and company’s Fred Reichheld. Reichheld, in a series of books like “The Loyalty Effect” (1996) and “Loyalty Rules” (2001) has demonstrated that there is a simple way of measuring the public attitude to a company. You simply ask a consumer:

“Would you recommend this product or service to a friend?”

If the answer is yes, the person is a “promoter,” if no, the person is a detractor. You deduct the first number from the second to reach something known as “the net promoters index.”

The index really comes into its own when you compare two or more companies in the same sector. The higher the positive number applied to a company, the more likely the company is to grow (more recommendations=more customers). The lower the number, the more likely the company is to shrink (more detractors=fewer repeat purchases).

Recommendations matter. In Emanuel Rosen’s book, the Anatomy of Buzz he points out that

• 65% of Palm customers heard about the device from another person.

• Friends and relatives are the number one source for information about places to visit or about flights, hotels or rental cars.

Of people surveyed by the Travel Industry Association:

• 43% cited friends and family as a source for information.

• 57% of customers of one car dealership in California learned about the dealership from another person

“This is not unusual,” says Jim Callahan of the research company Dohring.

[Source: James Torio’s “Blogging a Global Conversation” ]

The recommendations we would formerly seek from friends and family we increasingly find through Google. Research shows that 75% of shoppers use search engines with the intention of purchasing products and services . What they find when they get there is increasingly determined by bloggers and contributors to online forums. Recent research by Market Sentinel in the UK showed that, of Nielsen’s Top 50 grocery brands, 40% had negative commentary in the top ten results on Google UK . Searchers will even put keywords into searches to look for issues. For example, in the last 30 days, one of the popular searches tracked by UK-based paid search vendor has been “Ford Focus problems”. This is a search designed to flush out online resources dealing with customer service issues.

Why does this “net promoters” index matter? Why is it important to keep the number of detractors down, and the number of promoters high? According to Reichheld’s exhaustive studies, it is very accurate leading indicator of stock price. A high value means that your stock will rise, and a low value spells trouble.

I was reminded of this recently in the context of some work Market Sentinel and its partners are doing looking at the issues a famous corporation has been experiencing with its customer services function. The corporation, faced with a highly competitive marketplace, took the decision to offshore much of its customer services function to India. It cut down on expensive home visits which were an important constituent of its value add. Wall Street applauded and the stock price rose. However the number of negative comments about the company in message boards and blogs rose steeply. Customers did not appreciate the long waits for customer service. They did not appreciate the failure of the company to fulfil on its after sales promises.

Finally, earlier this summer, the company was targeted by a famous blogger who had bought one of its products and was frustrated not just by the poor quality of the product, but by the company’s failure to fulfil its customer service promise. The blogger started writing about it and was immediately deluged with comments and trackbacks by aggrieved fellow customers.

A couple of weeks back the company published its latest numbers. Growth has stalled (more detractors=fewer repeat purchases). The stock price fell like a stone. As Mr. Reichheld puts it in a paper published by the Harvard Business Review, the “net promoters” index is /”The One Number You Need to Grow”.

Market Sentinel’s benchmarking service provides an index of “net promoters online” and tracks it live. It is now our most popular service. Tracking corporate reputation online is the business equivalent of the Nielsen rating. And it is something managers and corporate investors simply have to do if they want to look at whether a company is on the way up, or the way down.