B2B or C, the answer should always involve Creative Publicity.

Marketing has long been concerned with finding differences to create a sustainable competitive advantage. As a result, lessons from consumer marketing have often been discarded as irrelevant for industrial marketing. This paper posits that the similarities between business-to-consumer (B2C) and business-to-business (B2B) matter more than any differences we can find. This article summarises and reinforces prior work with our similar findings, and the implications for marketers. This article was first published in April 2021 on WARC.com

===

The discipline of marketing must at least be 10,000 years old. Ever since the earliest farmers settled it enabled specialisation, surplus, and the resulting trade with its central question: how can we sell more to more people?

Producers have always tried to build reputations. Some of the earliest practices are to be found on amphoras used in Mediterranean trade as far back as 1500 BCE. Advances in welding and the availability of different steel types around the 9th?century led to the production of the then famous Ulfberht swords, which were of superior quality and highly sought after by those in the business of combat. They were the starting point of a medieval tradition of blade inscriptions – another early example of branding - and the equivalent of a Mont Blanc pen or MacBook Pro for today’s office warriors.

In the last two hundred years mass production and its standardisation made it even more important to be able to distribute this surplus beyond our regional communities, and eventually globally. Progressively our mindset has shifted towards reaching more and larger markets to grow, even to the idea of?blitzscaling, which, according to VC entrepreneur and co-founder of LinkedIn, Reid Hoffman, is: “ (...) the science and art of rapidly building out a company to serve a large and usually global market, with the goal of becoming the first mover at scale”.

So, fundamentally, not a lot has changed in the practice of marketing, perhaps other than the pace at which we expect to see results.

Organised marketing thinking on how to generate sales is a much more recent phenomenon. The first ever marketing course was allegedly taught at the University of Michigan by Edward David Jones in 1902. By the 1920s the marketing discipline was already dividing into several schools of thought. However, since the 1950s the ideas of micro-marketing by Alderson (1957) and McCarthy’s 4P’s (1960), accompanied by segmentation, targeting and positioning championed by the likes of Kotler (1967) and Ries and Trout (1969) have dominated marketing teaching. More recently it has started to incorporate other fields of knowledge, especially psychology, neuroscience and, as data has become much more available, data science with tools such as Artificial Intelligence and Machine Learning.

Ubiquitous computing power means we now have a colossal amount of data, not to mention a huge range of analytic methods available to us. This has an interesting effect, for, a bit like a cat and open doors, we cannot resist going through the door, every door, looking at every bit of the data. More data, more detail, more looking at the micro, more ‘richness’ without asking if it is a distraction or worse, the wrong road, as we lose sight of context and the bigger picture, the overall patterns. When we become primarily interested in differences we will find those by the truckload because they are there, albeit most will be tiny. But we will only know they are tiny by looking at the big picture, the overall patterns.

Same or different?

Source: Flickr, 'identical twins' by Diane Arbus at MOCA

The most dominant school of thought in marketing has been one that strongly advocates understanding how buyers - and businesses - differ from each other so as to be able to leverage those differences for advantage. In this paradigm, to assume that marketing to consumers as opposed to businesses are very different does not seem to be outrageous. However, we should also look at it the other way around, are marketing to consumers and businesses mainly similar?

Sure, industrial buying will differ from consumer buying in some respects. Typically, the absolute number of buyers in a given market is much smaller; often, the buyer and seller will directly engage in a process that can take up several months, even years, before a purchase is made; decision making is considered to be more rational. There can also be very specific contractual arrangements in place to secure supply, for example, which may lead to different purchasing patterns. We could go on. Nonetheless, we should look at the similarities. Work has been done on this and the first results were published several decades ago.

Around the same time, in the mid 1950s, when Wroe Alderson led the shift towards niche-marketing that still dominates marketing thought today, another marketing scholar started his own revolution which until recently has gone somewhat unnoticed. This is unfortunate as this work provides a great wealth of knowledge and evidence on how the discipline of marketing, regardless of to whom you are selling, should concern itself with a few fundamental truths. If so, the similarities between B2C and B2B might be much more significant than we initially anticipate. Notwithstanding all the schools of thought on how and why consumers or customers buy, Andrew Ehrenberg, a statistician, diligently set out to collect evidence about buyer behaviour which would challenge much of marketing thinking then and now. He took an unorthodox, even radical path by collecting and analysing purchasing data?before?positing a theory on why it manifested itself as it did. What he found were robust and reliable patterns which he was able to model. His first paper on this topic, published in 1959, was titled The Pattern of Consumer Purchases. In it he demonstrated that buying frequencies followed a Negative Binomial Distribution (NBD). Together with Gerald Goodhardt and Colleen Chatfield, he later refined his findings into a stochastic model of choice called the NBD-Dirichlet (generally known as simply, the NBD).

Below is a chart of what an NBD looks like for a big, medium-sized and small brand. In each case, the brand’s customer base is made up of a lot of light buyers. That means that when a small brand grows it successfully acquires new customers, the shape of its customer base will not change much. It will still be an NBD because most of its new buyers will be light buyers, with a few heavier buyers. It is one of the manifestations of the NBD model which shows that the main difference between big and small brands is in their penetration. Much less so in loyalty-related metrics such as purchase frequency or weight of purchase. Categories differ in the underlying relationship of penetration to loyalty measures, some are just bought more or less often, but the pattern is always that big brands command (somewhat) higher loyalty.

Sources: The Commercial Works client data (top and left), and Sharp & Romaniuk (2016)

Seventy years of research – of which a large part stems from the Ehrenberg-Bass Institute and its forerunners – found nothing but purchase patterns that closely fit the NBD – with only a few fairly obvious exceptions (e.g. usage, distribution and pricing effects). In the two scatter plots above, one can easily see the strong relationship between penetration and purchase frequency, where brands mostly differ in penetration and to a much lesser degree in their rate of purchasing, the effect also known as Double Jeopardy. For a long time, Andrew Ehrenberg worked at Attwood, now Kantar, analysing datasets mainly from CPG categories. So far, most of the work of testing the NBD has been in CPG, simply reflecting the availability of data, although the body of evidence from industries such as services and durables has considerably grown over time. Many of these cases are described in the seminal work of Byron Sharp, Jenni Romaniuk and their many colleagues in the book How Brands Grow (parts 1 and 2) which have led to a to a much wider appreciation and adoption of Ehrenberg’s pioneering work.

What to do about these findings? Ehrenberg advocated brands should be well-known, easy to buy and thought worth it – a theory of growth helpfully characterised by Sharp and colleagues as building mental and physical availability. This idea built on the view that a critical implication for managers was to appreciate the importance of creatively publicising the brand. Creativity mattered, even to a statistician. We’ll get back to the importance of creativity after looking at how relevant these patterns are for B2B marketers.

One might expect much rejoicing with so much of this evidence available to us. Now managers can track and benchmark performance, even make predictions on where future sales will come from. However, many still feel their situation or category is different. While there is much evidence that these patterns hold in B2C markets, surely, selling to businesses is quite different?

Admittedly, the number of studies in industrial or B2B marketing is considerably smaller. That said, let’s take off our differentiation glasses for a moment and check if we aren’t overlooking the fundamental similarities that govern results.

The NBD model has been proven to provide a good fit with purchasing behaviour in a great number of categories across products and services for thousands of brands over the course of many decades. Looking at the research in industrial marketing we actually also find a wide selection of categories with seemingly very different dynamics and large differences in their rate and/or weight of purchasing. From the mid-seventies onward, we now have evidence that contracts for aviation fuel, office products, residential courses, foreign exchange contracts, several professional services (media buying, marketing promotions, legal, accounting), medicine prescriptions, plastics, concrete, medical supplies and financial services follow the same patterns, and demonstrate a good fit with the NBD.

Below is an example from where purchasing is indeed very different from grocery shopping: civil aircraft buying. Each purchase involves 10s or 100s of millions of dollars, takes years, and will be sold, negotiated and scrutinised closely multiple people during the process. Despite these differences, the patterns in buying behaviour are remarkably similar: brands sharing customers in line with size, no niche brands with higher loyalty.

Source: Bennet (2016) – the table shows the actual data (Observed) against the NBD-model’s predicted values (Model).

Some might argue that some of these studies are too old (“just look how the world has changed!”). I would argue this was probably the same argument people gave in 1969 in response to the original 1959 paper. And from the people commenting in ‘79 on the data from ‘69, and... etc. I will highlight one other example from a very different B2B category to demonstrate how universal these patterns are, and to reassure people we can still rely on these principles today in order to build our brands for tomorrow.

Below is an example from our work in services. It shows data collected over four years from a globally operating business which sells services to other professionals. Total revenues exceeded several billions of dollars. Here, again, yet more pronounced than typically seen in the more well-known CPG examples, the reliance on light buyers is also instantly clear, the lightest buyers make up roughly half of the total revenue.

Source: The Commercial Works client data.

Clearly there are a few buyers that spend a lot (tens of millions a time), yet they are very few in number. A lot of resources are spent on keeping those clients happy. However, and given what we know about brand growth, it warrants a question about the balance being struck. We might get lucky and know or find one or two more customers who will buy at the high levels, towards the right-hand end of the chart (and if we can, we should) but it would be hard to deny that growth will come by adding all types of buyers to the customer base – and they will disproportionately be lighter buyers who tend to spend less. Do our strategies and tactics reflect this sufficiently? Or are we being (un)realistic in our expectations when it comes to our marketing targets in terms of revenue per customer?

Another interesting aspect of this chart is the extremity of the skew or reliance on light buyers. As noted by Wilkinson et al. in their study from 2016 on the fit of the NBD in industrial buying, some business-to-business markets might simply have a much larger pool of (very) light buyers compared to business-to-consumer markets. For example, there might just be a few companies spending large amounts. The chart above seems to suggest this is also the case here.

Selling direct to end users will allow some B2B businesses to check whether there is an NBD in their own data. Compiling an image that includes data from competitors to demonstrate the level of switching and compare loyalty levels might be more difficult but is not impossible. In some cases, the data will look rather patchy as there are fewer transactions which might make the numbers ‘jump around’ from period to period. One obvious solution would then be to measure over a longer period of time, and the patterns will likely emerge (as was the case in the example above).

The study from Wilkinson et al. is interesting for another reason – especially for those working in B2B industries. It presents the use of Conditional Trend Analysis – an extension to the NBD introduced by Ehrenberg and Goodhardt in 1967 – as a means to determine and benchmark from which buyers changes in sales are expected to come from. In other words: how likely are certain types of buyers (e.g. non-buyers) to buy in the next period?

Because B2B marketers often deal with a smaller pool of buyers (and often know who they are), it might allow them to better pinpoint sources for growth and resource allocation. Some may want to revisit their Account Based Management (ABM) strategies on the basis of these analyses.

If we have limited data we should ask ourselves a fundamental question: is our situation really that different? Because if you embrace the idea of the unchanging and universal nature of our buying, regardless of what you’re selling, it opens up a wealth of similar examples you can use to improve the effectiveness of your marketing. We should perhaps resist that first urge to instantly jump into differentiation mode. We should still be sceptical, just as much as we should stay curious and try to find out if these patterns hold in one’s own respective industry or situation. But, fundamentally, it is about people and how they choose, even in committees and with complex net value measures. Which is why we should look carefully at the similarities as well as take note of the differences. As is often the case, it is all about balance. The balance between acknowledging and having an interest in the individual case versus relying on well-established knowledge.

This article has summarised the most fundamental patterns in buying behaviour, and how they also manifest themselves in B2B categories. Alongside the dominant and more widely taught high-engagement model of marketing, we have a substantial, evidence-based alternative: the low-engagement model of marketing. This model predicates brands will mainly grow by increasing the size of their customer base, attracting mostly (very) light buyers and encourages managers to target all category buyers as the sharing of customers will largely be in line with size – not particular segments or brands. Most marketing will be about maintaining your rightful place in peoples’ minds, on store shelves, contact lists and the web.

When WARC asked me to write something on this topic, my first thought was: “don’t we know all this by now?” Perhaps not. In 2019 Adobe’s CEO stated that ‘retention is the new growth’. More recently, Gartner published their Gartner CMO Strategic Priorities Survey 2021 in which it surveyed close to 400 CMOs globally from numerous industries. Apparently, 73% of them indicated to rely on existing customers to fuel growth in 2021.

Therefore, there likely isn’t any harm in re-iterating the work of others as well as offering up new examples that confirm what we’ve known for a very long time. After all, we should ourselves appreciate the low-engagement model of marketing: people are busy and tend to forget rather quickly. To that extent, this paper complements two recent studies published by the LinkedIn B2B Institute. In the first, Les Binet and Peter Field specifically analysed B2B campaigns from the IPA databank. They acknowledge their dataset is limited but they find no reason to believe marketing is fundamentally very different. The second study, conducted by research agency System 1, echoes the same sentiment: it finds that B2B still has a lot to learn from B2C. Three quarters of the 1600 B2B ads tested apparently failed to contribute to what they consider drives long-term share growth. These ads relied more on outspending the category versus leveraging the strength of the creative – which helps build memory structures, improves recognisability and evokes an emotional response.

It’s all about people buying and people don’t change much. We are still working with Homo Sapiens version 1.0; a cognitive miser, emotionally dominated though rationally capable (in short bursts) and forgetful. This crowded mind needs reminding, steadily, whether it’s about your industrial ducting, electricity supply or chocolate bars. The easier it is to get a place in that memory, the more chance you have of growing, one light buyer at a time, a few of whom will become a bit more loyal, but only for a while, but that’s another story.

Selling and buying have been going on for a very long time and it appears we can rely on the universal, unchanging nature of buying - and exploit it to our advantage. Perhaps we should worry less about whether we’re marketing to a B or a C, because the answer is always better when it includes good quality creative publicity.

Sources

Bennett, DR (2016). An empirical study of industrial consumer buying behaviour: how airlines buy airplanes.?45th Annual Conference of the European Marketing Academy (EMAC 2016).?Oslo, Norway 24 - 27 May 2016

Bhattacharya, C. B. (1997), “Is your brand’s loyalty too much, too little, or just right? Explaining deviations in loyalty from the Dirichlet norm”,?International Journal of Research in Marketing, Vol. 14 No. 5, pp. 421-435.

Ehrenberg, A, (1959) "The Pattern of Consumer Purchases," Applied Statistics 8, 1: 26-41.

Ehrenberg, A., Barnard, N., Kennedy, R. and Bloom, H. (2002) Brand Advertising as Creative Publicity Journal of Advertising Research, July-August 2002, 7-18.

Ehrenberg A.S.C. Uncles M.D. Goodhardt G.G. (2004) Understanding brand performance measures: Using Dirichlet benchmarks. Journal of Business Research; 57 (12): 1307- 1325.

Graham, C, Bennet, DR, Franke, K, Henfrey, C & Nagy-Hamada, M. (2017). Double Jeopardy – 50 years on. Reviving a forgotten tool that still predicts brand loyalty. Australasian Marketing Journal. 25 (4), pp. 278-287.

Goodhardt, Gerald J., Andrew S.C. Ehrenberg, and Christopher Chatfield (1984), "The Dirichlet: A Comprehensive Model of Buying Behaviour," Journal of the Royal Statistical Society, 147 (part 5), 621- 55.

Goodhardt, G. J. and Ehrenberg, A. S. C. (1967), “Conditional Trend Analysis: A Breakdown by Initial Purchasing Level”,?Journal of Marketing Research (JMR), Vol. 4 No. 2, pp. 155-161.

Romaniuk, J. and Sharp, B., (2012), How brands grow part 2. South Melbourne: Oxford University Press. Sharp, B., (2010), How Brands Grow: What Marketers Don’t Know. South Melbourne: Oxford University Press.

Sharp B., Wright M. & Goodhardt G., (2002) Purchase loyalty is polarised into either repertoire or subscription patterns. Australasian Marketing Journal; 10 (3): 7-20.

Snijders, W. and Graham, C. (2018) Majority Report, Eat Your Greens: Fact-Based Thinking To Improve Your Brand’s Health, APG Publishing

Uncles, M., A. S. C. Ehrenberg, and Hammond, K., (1995) “Patterns of Buyer Behaviour: Regularities, Models, and Extensions.” Marketing Science 14, 3-2: G61–G70.

Wilkinson, J.W., Trinh, G., Lee, R., Brown, N., (2016), Can the negative binomial distribution predict industrial purchases? Journal of Business & Industrial Marketing 31 (4), 543-552

https://www.adnews.com.au/news/retention-is-the-new-growth-says-adobe-ceo

https://business.linkedin.com/content/dam/me/business/en-us/amp/marketing-solutions/images/lms- b2b-institute/pdf/LIN_B2B-Marketing-Report-Digital-v02.pdf

https://www.marketingweek.com/majority-b2b-creative-ineffective/

https://www.gartner.com/en/newsroom/press-releases/2021-01-19-gartner-survey-shows-73--of-cmos- will-fall-back-on-lo