Aug 14, 2024 | Analytics, Marketing, Sales
There is lots of talk, often sales-hype from digital urgers, about Lifetime Customer Value. When applied correctly, it is a vital measure, but when you look closely, it often means lifetime customer revenue.
Revenue is of little commercial value in the absence of margin, so the discussion can be completely misleading.
Understanding the margin generated by customer segments, or in some cases, individual customers is an immensely valuable metric that enables you to focus resources where there is the most benefit to the enterprise. You can make informed tactical choices with a great level of confidence based on the margin delivered.
Customer margin is also an enormously useful metric elsewhere.
Sales people are often rewarded on revenue, which can be gamed. Margin over time is much harder to game, and a far better measure of the effectiveness of a salesperson in delivering value to the enterprise while serving customers.
Similarly, calculating the cost of acquisition of a customer gains traction when measured against margin rather than revenue.
One of my clients businesses relies on referrals as a source of business. Increasingly they are moving towards margin on converted referrals as the single metric that best measures the impact of their marketing and product delivery efforts.
You cannot generate margin in the absence of revenue, but you are easily able to generate revenue without margin. Not a good idea!!
As an aside, also beware of the difference between margin and mark-up. They are similarly often used to mislead the unwary.
Jun 11, 2024 | Analytics, Governance
Goodhart’s law is a much quoted adage that states: When a measure becomes a target, it ceases to be a good measure’.
When we see numbers cited as evidence, we tend to instinctively give them more credibility than they may deserve. Without an examination of the source and scope of the numbers, just believing them on face value can lead to very bad choices.
Charles Goodhart is a prominent British monetary economist. His public profile started as a footnote to a 1975 article and read: ‘any observed statistical regularity will tend to collapse once pressure is placed upon it for control purposes’. It took on its now well-known form, above, when restated by an anthropologist seeking to link the idea to a much wider context than economics.
In almost every business I visit, I see examples of Goodhart’s law. The most common is the use of EBIT as a measure of success. The reality however, is that it is simply a financial measure of the outcome on the myriads of smaller actions and decisions that have been taken in the deployment of resources.
Two of the most public failures of measurement should make us all wary of using quantitative targets as a measure of performance.
The 2010 Volkswagen software scam. VW installed software in a range of cars with their turbocharged direct injection diesel engines that ensured that when being tested, the cars made the emissions standard of the US EPA.
The Vietnam War body count. The US defence department sought to justify the billions of dollars and thousands of lives expended in the defence of a corrupt government in South Vietnam. The measure became the body count of VC fighters and North Vietnamese regulars, which led to wholesale slaughter. These imaginary and bloody numbers were used for years as the basis for increasing commitment. It finally became obvious that they did not in any way reflect the ability or willingness of much of the population to fight what they saw as aggression by the US.
There are thousands of examples. If you looked, you would see them every day, especially when politicians open their mouths and quote numbers.
Header Photo: Charles Goodhart. Professor Emeritus at the London School of economics.
PS. When you look at the header, Dr. Goodhart is looking directly into your eyes. You will tend to believe anything he tells you.
Apr 10, 2024 | Analytics, Communication
It seems that ‘the truth’ is a malleable concept.
We are overwhelmed by opinion masquerading as fact, economic and social models designed to deliver a predetermined outcome, managed correlation equated to causation, and market research that asks the wrong questions of the wrong people.
What is truth to one person is nonsense to another.
We should be able to see ‘the truth’ about what has passed, there is data that should distinguish fact from fiction. However, we still fail to discern the truth from amongst the data available for analysis.
Who is winning the war in the Ukraine?
Depends on who you ask, and both sides have data that shows conclusively that they are winning.
Remember Vietnam? I do.
The Americans had an overwhelming advantage in material, technology, and logistics. How could a little country with few resources and no technology of their own, face and win against the mightiest war machine the world has ever seen?
Impossible but it happened.
Until the Tet offensive commenced in January 1968, there was no doubt in anyone’s mind, apart from the North Vietnamese, that it was only a matter of time until the might of the Americans became overwhelming.
The Americans had data that proved to them they were winning, despite the secret conclusions contained in the Pentagon Papers. It was not until the spring offensive in 1974 that it was obvious to all that the American ‘Facts’ that were being analysed were irrelevant, and the conclusions drawn were terminally wrong.
The clear answer to the question in the header is: ‘only when you analyse the right data.’
Header credit: Hugh McLeod at Gapingvoid.com
Mar 15, 2024 | Analytics, Marketing
Every market research proposal must answer a duo of critical questions before it proceeds, if it is to be of any value.
What is it for, and how will it be used?
Market research is done for all sorts of reasons. Many commissioned projects have little to do with the examination of the critical factors in driving success.
They just provide a convenient crutch.
Several projects commissioned and paid for from marketing budgets I controlled would come in under the ‘what the F&&k’ category. However, in my defence they were usually quant studies designed to generate the numbers necessary to pass the accountants various thresholds. This enabled me to progress projects that qualitatively and ‘in my guts’ were winners. That is the way they usually turned out!
In the absence of clearly understanding how the research results were to be used, how they would add strategic, operational, or technical value, why should you bother?
There is a further tier of understanding that is required: Are you looking to define an objective outcome, or are you seeking understanding and insight?
In the case of the outcome required being quantitative, simple yes/no, black/white answers to a question are sufficient.
When you are looking for insight, there may be a few numbers, way below a level of statistical significance, but they can be reassuring. However, the value lies in discovering the connections, implications, options, and potentially hard to anticipate consequences.
Research is a critical step in successful marketing programs. However, in the absence of a very clear and compelling answer to the ‘What is it for’ question, it should not proceed.
The header illustration is the only AI used in this post.
Feb 7, 2024 | Analytics, Marketing
Every customer segmentation exercise I have ever seen is based on geography, demographics, some combination of behavioural characteristics, or all of the foregoing.
‘Young women, 25-35, single, who live in the Eastern suburbs, earn more than 80k, and eat out a lot’ sort of analysis.
Misses the point.
There are five types of customers in every business I have ever seen
Unhappy. These will often tell you and anyone else they can grab, of their unhappiness. Usually these are users, rather than the ones who make the purchase choice. This means they can be a fantastic source of improvement ideas, but can also consume lot of your time with things that cannot be changed.
Satisfied. When a customer is satisfied, they go away happy and you rarely hear from them. The more time you spend understanding the drivers of their satisfaction, and doubling down on them, the better.
Loyal. This group of people usually quite small will not go anywhere else and will generally pay premium to you in the knowledge that you will not fail them. In effect, it is in effect a risk mitigation strategy for them.
Apostles. Apostle customers these are generally small subsection of your loyal customers and occasionally just a satisfied customer when conditions are right who are prepared to aggressively push your case to others in their various networks. These people are your best salesman and also your cheapest, although there is a cost get him to getting them to the point where they will proselytise on your behalf
Cheapskates. The fifth type, the one you can probably do without, is the one who dips in and out of your product, chasing the cheapest price irrespective of other considerations. It also seems to me from experience, that they are also the ones who complain a lot.
Think about it.
I am prepared to bet there will be nuggets of value hiding in plain sight you can use.
Header credit: My thanks to the exiled Scott Adams, and sidekick, Dilbert.
Jan 17, 2024 | Analytics, Strategy
Imagine. Possibilities.
‘Strategic thinking’ has been overtaken by the ‘quants’.
Those that believe that by generating loads of data, analysing past events, behaviour, and outcomes, you can create a model that will give answers to the key strategic question: How best to deploy limited assets for the best return’?
Aristotle 2,500 years ago observed that in some things the past will always be the same as the future. Think about gravity. We know it will be there tomorrow exactly as it is today.
Your task in this case is to identify and quantify cause and effect.
Aristotle also observed that in other things it is not the case that what happened yesterday will be repeated today. In that case, you must form hypotheses, test them, learn, then rinse and repeat.
In other words, you need to imagine possibilities.
Look at the evolution on the mobile phone for evidence. On January 9, 2007, Steve Jobs officially announced the original iPhone. On January 10, 2007, despite luminaries like Steve Ballmer poking fun at it, all preconceptions about what a mobile phone was, were out the window. The past was not representative of what the future would look like.
The world is a messy place, today rarely looks like yesterday. In that messy place our task is not to look at the past and project onto the future, our task is to imagine possibilities.
Strategy development is all about imagining those possibilities, making choices on what appears to be the best bet, and putting your money down, adjusting as necessary as more information and insight are gathered.
Aristotle did not conceive the OODA loop. He left that to John ’40 second’ Boyd 2,500 years later, but it was inherent in the ‘scientific method’ he articulated, and should be required learning for every decision-maker.
Header is a representation of the ‘Johari Window’, made famous by Donald Rumsfeld
