Aug 7, 2026 | Analytics, Marketing
Twelve years ago, I argued that your website was your digital home. You could run around naked if you wished. Your home, your rules.
Social platforms offered rented accommodation. The landlord made the rules and could change them without asking you. However, it was not clear to me then that it would become an existential distinction.
The traffic coming from these sites was never ‘yours’. It was borrowed from the platforms and subject to their rules, algorithms, and monetisation strategies. The landlord has struck and imposed a pile of new requirements and limitations on your rented digital channel access.
The emergence of AI has been the pivot point.
Search is no longer an appropriate term. It implies looking in several places, considering options, enabling you to make informed choices. The overviews now being delivered give what the platform regards as the default best answer, above any alternative potential source.
Most of us take that answer for granted and look no further, most of the time. Google, and increasingly all platforms, are providing disincentives to leave their ecosystem, while promising incentives by way of algorithmic distribution to stay.
As a result, traffic to websites via digital platforms has crashed.
In the case of the modest StrategyAudit blog, the traffic referred by Google and LinkedIn, the major channels I used has eroded by around 80% over the last 3 years, accelerating noticeably in late 2024 when Google rolled out overviews to Australia.
LinkedIn has progressively rewarded content that keeps users inside LinkedIn and restricted the reach of posts that send them elsewhere. This week saw another turn of the screw. The garden wall had grown higher again. I received a note from Microsoft telling me that Publisher will be removed from the Microsoft 365 subscription from October 1. The opening sentence of the email is ‘Microsoft is committed to improving your Microsoft 365 subscription’
How removing something I have paid for and is of value to me is an improvement in the absence of a price reduction is beyond my simple understanding. It does however make the point that the landlord can do whatever they like and you have no redress. They very kindly gave me list of instructions about how best to preserve publisher files so I could use them again. This is a bit like the landlord putting on a sad face and giving you a list of alternative rented locations as they throw your furniture out onto the street.
We gave the digital landlords our attention, data, and content. In return, they gave us apparently free access. They have not broken that bargain so much as rewritten it, without negotiation.
For those like me who use the web as a creative outlet with no expectation of a financial return, the only impact such changes have is on my ego. So few people are reading and responding to my brain-farts. However, for someone who had built a business on rented digital space, it is a rapidly advancing disaster, as the foundations of their business model are being rapidly removed.
The only antidote is to build your own digital house and fill it with furniture valuable to those with whom you wish to interact, so they choose to visit. The platforms must become the paths that lead to your home, not the real estate upon which your home is built.
P.S. The link in the first sentence of this post sends me to ‘LinkedIn gaol’. You will be lucky to find this post on any platform. Subscription, which is absolutely free and will not solicit anything, is the only way to ensure you see future posts.
Aug 4, 2026 | Analytics, Marketing
Calculating probabilities of an outcome is not the same as making a prediction.
What is the probability that a potential customer will buy your product rather than an alternative?
What is the probability that they will choose one SKU rather than another from your range?
Will your new advertising tagline remain in their memory long enough to influence a purchase?
Will a price increase reduce volume by more than the additional margin generated?
These are probability questions, but that does not make them the same statistical problem. They are not predictions.
Marketers have always tried to put numbers around questions like these, because the accountants and engineers who generally run the place distrust anything they cannot squeeze into a spreadsheet. A number, particularly one that looks like a detailed calculation, attracts far less scrutiny than an honest admission of uncertainty.
Probability models cannot remove uncertainty. They can, however, show us what normal buying behaviour looks like, and provide a benchmark against which we can test our assumptions.
Two drivers in any repeat purchase market
In established FMCG and other repeat-purchase markets, two forces largely shape buying behaviour.
The first is how often someone buys from the category.
Take the yoghurt market as an example. In every case, the behaviour and volume of buying will vary from heavy users to occasional buyers, from brand and sub category agnostic price buyers to brand advocates, and every point between.
Statisticians describe the pattern created by these different buying rates with the eye watering name of Negative Binomial Distribution, usually shortened to NBD.
NBD does not tell us that an individual will buy yoghurt next Thursday. It describes how purchase frequency spreads across the whole population: a few heavy buyers, many light buyers and a group who buy nothing during the measurement period.
The second force is brand choice.
Most buyers choose from a brand and variety repertoire. One brand may dominate their purchases, but price, availability, flavour, pantry stock and the occasionally volatile demands of the household influence each decision.
This is the Dirichlet model which describes how buyers divide their purchases across that repertoire. It is a weighted average of buyer behaviour across the market for each individual possible choice and combination of choices.
Think of each buyer rolling a set of weighted dice. Every brand appears on the dice, but some brands occupy more faces than others. The result remains uncertain, but it does not remain completely random.
Combine category purchase frequency with brand-choice probabilities and you can build a picture of the market. Statisticians would call it an NBD–Dirichlet market model.
From untidy households to stable market models
Continuing the yoghurt example.
The market contains multiple brands and many SKUs covering plain, fruit, Greek, low-fat, full-fat, lactose-free and some emerging and specialty products that represent a purchase choice in the wider yoghurt market.
Each household behaves differently. One buys frequently and moves between several brands. Another buys occasionally and nearly always chooses the same product. A third selects whichever brand carries the discount sticker.
Individual purchases look erratic. Add thousands of them together over time, and recognisable patterns emerge.
Large brands usually win because more people buy them, not because their customers display dramatically greater loyalty.
Smaller brands suffer a form of ‘double jeopardy’. They attract fewer buyers, and those buyers tend to purchase them slightly less often.
Buyers also share their purchases across competing brands. Your customers do not belong to you. They belong to the category and sometimes, when it suits them, buy your product.
Having operated in many FMCG categories over the years, those observations have held true in every case.
A model is not a crystal ball
The NBD–Dirichlet model works best in relatively stable markets where customers make repeat purchases and treat the competing brands as reasonable substitutes.
Defining the boundaries of the market is therefore crucial.
A parent may not see a child’s yoghurt pouch and a tub of plain Greek yoghurt as alternatives. Combining them in one model will produce statistical anomalies. The analysis should separate meaningful subcategories where buyer behaviour shows clear partitions.
The model cannot tell you whether a new tagline will lodge in buyers’ memories. That requires creative testing and evidence of memory and behavioural effects.
It cannot calculate the price elasticity, the contribution on margins of price changes, and likely competitor responses.
It cannot reliably forecast a genuinely new category for which no pattern yet exists. That task remains in the hands of the creative marketer, an increasingly valuable person in this age of ‘AI everything’
Importantly, the model cannot explain why every change occurred. A promotion, stockout, new distribution agreement, competitor withdrawal or advertising campaign may shift the observed probabilities. The model provides the baseline that helps us recognise when something unusual has happened.
Marketers should use probability models to challenge assumptions, establish realistic benchmarks and identify deviations and outliers worth investigating.
They should not be used as a crutch for decision making, as they cannot tell you why change has occurred.
Note: the combination of NBD and the Dirichlet models to reflect behaviour in a market comes from the Ehrenberg-Bass institute for marketing science.
Jul 30, 2026 | Analytics, Management
Assembling and leveraging the complex mix of information and capabilities required to ‘win’ is a challenge similar to climbing a slippery ladder.
Most rely on data, spreadsheets, and extrapolations, blended in with optimism.
At the bottom of this ladder is data. Raw, unfiltered, and almost useless on its own. Just numbers, words, events. Noise without context. Like a kids jigsaw puzzle upended in a pile on the floor.
The ‘ladder’ has five steps.
Data. The pile of numbers, littered around files, outside sources of information, and peoples brains, without context relevant to the questions being posed.
Information. You start sorting. Cleaning. Labelling. Structure begins to emerge. It’s janitorial, unglamorous, tedious, and essential. This is where many people mistake motion for progress.
Knowledge. Connecting dots delivers knowledge. You see cause and effect, frequency, variation, and importantly the outliers. You can answer questions like “What’s happening?” and “How often?” Knowledge is a map, but it is still two-dimensional.
Analysis. This requires you test the knowledge, isolate variables, run comparisons, identify the outcome drivers, and develop ‘what if’ models. You don’t just have answers; you know how strong the answers are, and where they might fall apart. It gives you that added dimension.
Insight. Here’s where the real game begins. Insight is not just about seeing patterns, it’s about seeing meaning. It’s the inflection point, where a thousand observations converge into a single, clarifying idea.
But insight alone is not enough.
Running parallel to this clean, rational ladder is a tangled, intuitive, and very slippery pole: wisdom. Hard to see, impossible to teach, and absolutely critical to success.
Wisdom doesn’t follow steps. It grows from pattern recognition forged from lived experience. Mistakes made, ‘hunches’ that turned out to be right, and often spectacularly wrong, opportunities misread or missed completely. It takes time, perseverance, an ability to learn, curiosity, and determination.
It’s what lets a seasoned operator say, “That’s not going to work,” without needing a single chart, and be right. It’s instinct layered over time, a mental shortcut engine built on scars, sweat, and sleepless nights. It’s experience distilled into reflex.
The irony is that the higher you climb the corporate hierarchy, the more you need that slippery pole of wisdom.
If you only trust the data ladder, you might build the wrong solution brilliantly. If you only trust the pole, you risk hubris.
But when you climb with both, methodical analysis on one hand, and wisdom on the other, you are better able to understand problems, see the wider context, and spot points of leverage.
That’s the difference between a strategist and everyone else.
Jul 20, 2026 | Analytics, Operations, Strategy
If a job that took eight hours now takes one, most managers will call that a productivity gain.
They are right up to a point.
Economists usually measure labour productivity as output per hour worked. If we produce the same output in one hour instead of eight, productivity has risen whether we use the remaining seven hours to win another customer, dissect the accounts, or head for the beach.
The beach does not cancel the efficiency gain, it simply exposes the limits of the measure.
The conventional measures tells us how efficiently we produced what it can already see. It tells us very little about what happened to the released capacity, whether the output improved, or whether technology enabled valuable work that nobody previously attempted.
It is a stopwatch with a blindfold.
This matters as businesses adopt AI. Most discussion starts with replacement: write the report faster, answer the email sooner, produce the first draft without paying someone to stare at a blank screen.
Useful, certainly but not transformative.
If AI cuts eight hours of work to one and the business fills the other seven with more of the same work, the gain will eventually appear in output, margin or capacity. If the business simply adds seven hours of meetings, the technology has delivered potential while management has squandered it.
That is not a measurement failure alone. It is a management failure with excuses.
The larger opportunity comes when AI extends what a business can do.
Most SMEs sit on piles of data they barely use. The accounting system contains patterns in customer profitability, purchasing, margin leakage and payment behaviour. The CRM contains clues about lead quality, objections, lost sales, response times and repeat purchases.
Managers have always suspected gold lies in those records, but lacked the time, skills or patience to dig it out.
AI changes the economics of the digging.
A business can now examine why apparently similar customers generate wildly different margins. It can compare the language used by prospects who bought with the language used by those who disappeared. It can identify which products create revenue but consume so much handling, discounting and after-sales attention that they destroy profit.
None of this necessarily replaces an existing task. Often, nobody did the task in the first place.
That makes “hours saved” a miserably small way to assess the value.
The better questions concern decisions improved, waste prevented, risks spotted and identified opportunities leveraged.
My commercial history includes a significant time driving sales and profitability of dairy products.
Take yoghurt for example. The supermarkets report a yoghurt market. That broad category helps retailers size the category shelf space, then allocate space to product groups, brands and individual products. It creates a ‘category map’ products and preparation of management reports. It conceals much of the behaviour that drives the sale to the time poor, budget limited, harassed by kids shopping consumer.
Shoppers choose between fruited and plain, low-fat and full-fat, pot-set and stirred, indulgence and health, adult and children’s formats, single serves and family tubs. Then they behave inconsistently, driven by the behaviours at home that result in consumption. Shoppers have never read the segmentation presentation, and do not care about it at all. The same persons trolley can contain Greek yoghurt for breakfast, a sweet pouch for a child, a set in pot natural yogurt for a fussy daughter, and a premium dessert for Saturday night, and be different on every visit to the supermarket.
Traditional analysis could reveal these patterns, given enough clean data, statistical skill and time. AI does not invent the possibility. It lowers the cost and effort required to pursue it, then allows a manager to question the data repeatedly rather than wait three weeks for another report.
They can uncover submarkets hidden inside the convenient fiction called “the market”. It can improve assortment, pricing, promotion and product development. The value comes not from producing the old market report faster, but from seeing choices the old report blurred out.
Then comes the hardest category: genuine innovation. Not the pack design change, or the ‘new improved’ standard product, genuine, category generating new products.
Every important general-purpose technology eventually enables products, services and business models that people could not sensibly describe at the beginning. We struggle to measure these gains because measurement follows observable activity. It cannot count a commercial possibility before somebody discovers it.
Nor do conventional accounts handle quality and unpriced value in anything but hypothetical terms. A better diagnosis, a more useful forecast, improved process, or a tailored customer answer may create considerable value without creating a measurable additional unit of output. We do not need to throw out existing productivity measures. They answer an important question and allow comparisons over time.
Businesses need a second dashboard.
Alongside output per hour, managers should track the use of released capacity, the valuable work newly performed, improvements in product quality and management choices, improved cycle and takt times, and new revenue or margin that technology made possible.
That is the productivity challenge in businesses.
AI will save time. This is an opportunity also available to competitors and especially new market entrants, unencumbered by past choices and technology. The sustainable advantage will go to the businesses that know what to do with the no added cost capacity that better use of time can deliver.
As an interim GM 20 years ago I was contracted to a manufacturing business that while marginally profitable, was struggling with capacity. Analysis revealed what everyone knew at some level, that a key step in the manufacturing process was acting as a bottleneck, strangling capacity. A number of relatively simple process changes that required no capital and no personnel changes enabled the key step to operate an extra 6 hours a day, effectively nearly doubling capacity. That resulted in shorter customer lead times, dramatic increases in DIFOT, abolition of outside warehouse costs, and very happy customers increasing orders at the expense of competitors. The changes took a couple of months to work their way through the financial system, but when it did, the impact on profitability was substantial. However, it was only with hindsight and deep digging into the numbers, that the changes in productivity numbers could be tracked.
The foregoing examines the productivity challenge of an individual business.
In the recent budget, and following discussions, the Treasurer has loudly and consistently pointed to the ‘productivity package’ contained in the budget papers.
The challenge in delivering on the promise, and having the electorate understand the complexity of the productivity challenge faced by the Australian economy is immense. It is easy to pick at the edges and observe ‘nothing to see here’ because the policy documents are shallow, predictably lacking in any real and measurable project timelines and checkpoints, and the press releases are as useless as the claims of a carnival hawker. It is however a start.
Feb 4, 2026 | Analytics
‘How much is my business worth’ is a common question I get.
There are as many ways to value a business as there are consultants willing to charge you for a calculation. Sensible people use a range of tools, all of which in one way or another, seek to quantify future cash flow. That is the only reason someone would buy a business: they can extract more value from the capital deployed buying it, than deploying it in other ways.
Tobin’s Q ratio is one of those common tools that will deliver a number that is worth consideration, along with the many others.
It is a financial metric that compares the market value of a company to the replacement cost of the assets of the company. It helps investors figure out if a company is overvalued, undervalued, or fairly priced.
In a time when the value of a business is significantly influenced by the valuation of intangibles, all those items that deliver value to a buyer, but which do not make their way into the financial statements, you also need to consider how these will be valued, which is an entirely separate exercise, and by far, the most contentious.
The simplest way of thinking about the Q ratio is to consider:
- Market value is what investors think the company is worth—this is the company’s stock price multiplied by the number of shares. That applies for a publicly traded business. It is much harder to calculate when there is nowhere that shares can be traded easily. In that case, the judgement becomes more subjective, taking the place of the sentiments of the market that determines the value of a publicly traded share.
- Replacement cost is what it would cost to replace the company’s assets. How much it would take to rebuild or replace everything the company owns, including the intangibles. It is relatively easy to get a valuation of the physical assets, but much harder to calculate the value of a brand, the experience in your employees heads, the strategic position in a market you hold, and the list of customers you have.
A ‘Q value’ greater than 1 means that the company’s market value is higher than the cost of replacing its assets. This could mean investors are optimistic about the company’s future growth, or the company might be overvalued.
A ‘Q value’ less than one means the company’s value is lower than the replacement cost of its assets. This could mean the company is undervalued, or investors don’t think the company will perform well in the future.
Obviously, when the Q value is 1, the market value is in balance with the replacement cost of the assets, including intangibles.
The really challenging bit is putting a value on the intangibles. For that you need deep marketing experience, domain knowledge, and wisdom.
Jan 15, 2026 | Analytics, Governance
Many of those who run SME’s turn off when their accountant starts mumbling about ROE and ROA at the annual $500/hour financial roundup.
Understanding the difference is important to the long term health of the enterprise. They reflect how effectively a company’s management team is doing its job of managing the capital entrusted to it.
The primary differentiator between ROE and ROA is the amount of debt compared to equity is being used. Together they are a measure of the efficiency of the enterprises efforts to generate profit.
The difference is that ROA takes into account debt, while ROE does not.
Return on equity (ROE) is net income divided by shareholder equity. It measures profitability by relating how well a company generates profit from money invested by shareholders.
Return on assets (ROA) is net income divided by total assets. It’s an efficiency measure of how well a company is using its assets.
The difference is important for several reasons.
- ROE is focusses on the return generated on the shareholders’ equity. The easy analogy is the interest you receive on the balance in your bank account.
- ROA is focussed on the return generated on the total assets of the company, including debt.
In the absence of debt, ROE and ROA would be the same.
Viable businesses are able to make choices about the financial leverage they apply to their operations. The mix of debt and equity they employ.
A business that no longer has that option, in other words, investors have better options to generate a return on their funds, they will be withdrawn. At that point it often becomes difficult to borrow to replace the withdrawn equity, so the business becomes progressively unable to fund operations. At that point, it requires a ‘restructure’, or goes into liquidation.
Leverage works both ways.
ROE is used by investors to assess the profitability relative to their investment.
ROA is used to assess management’s efficiency in using all the enterprises assets, both debt and equity to generate profits.
