Understandingtrading psychology is one of the most important but alsoless often done tasks for investors.Of course, everyone realises that they need to analyse the investments they are considering buying.But many traders do not realise that winning in investment is mostly about successfully predicting what other market players will do.Before they do it. And that is a psychological task.
Most of the advice on the internet is not really psychology.It is quasi-psychology.You might get famous traders telling you things like “I always played tennis in the morning before my best trades to make sure I felt good.”This is useless.By all means, study what these guys do. You may get insights into how they look at opportunities and maybe any tricks they have for bouncing back from a loss.But famous traders don’t have any specific training in psychology. So, if you are specifically wanting to improve your own trading psychology, adopting their tips (such as the tennis one above)won’t really help you in achieving that goal.
What you need to look at is the psychological literature. This is published in the academic journals. You also need to do that having spent a decade on the trading floor. Fortunately, you don’t need to do that yourself since I have done it for you. So all you need to do is read the book.
The other sort of person who cannot help you is someone merely described as a psychologist. That’s better than nothing. But they need to be the right sort of psychologist, as I will set out in the next section.
Is that the right sort of psychology?
There are some actual psychologists who write on the topic and are experts in the field.But be careful about their specialisms.Someone who is a clinical psychologist may be an expertin schizophrenia. They may however not necessarily know any other aspects of human psychology. And of course these expertsdo not have any serious trading experience. So they definitely can’t help you improve your trading psychology. Unless you have schizophrenia. But in that case, you have more pressing issues to attend to than the performance of your portfolio.
There are also a lot of completely unqualified people who write on these topics. They don’t know any psychology and they don’t know anything about trading. These individuals are all over the internet. But you should not waste any of your time on them.
To identify the right sort of person, you need to ask two questions. Does this person have significant trading experience? Are they qualified in a related field?I am one of these people.
How To Optimise Your Psychology
To convince you of this, I will outline my ideas on how to optimise your trading psychology.The first thing to know about is that we have a lot of cognitive biases. (Here’s a list on Wikipedia: there are hundreds! https://en.wikipedia.org/wiki/List_of_cognitive_biases )
These are mental shortcuts that are often useful when we want a quick and dirty answer. However, they are often very unhelpful when we are trying to get something right.One example is Confirmation Bias. This is where people look only for evidence that supports what they already believe.There have been manyrobust psychology experiments published,that show time and time again that we do this often.
This isn’t something which is optional. Cognitive biases are a fundamental part of our wiring. More intelligence does not make you immune to biases. At least if you know about them though you have a chance to counteract them.
The first thing to note here is that if you use this bias when making your own trading decisions, you will make bad decisions.Every time!So you will definitely not be optimising your trading psychology.But here’s the key point: everyone else in the markets will be doing it too.
So what does that mean?It means you need to know about Confirmation Bias. You need to think about it in a market context.Look out for it in yourself and be careful.Expect it in other market players and trade accordingly.
That’s how you stand the best chance of optimising your trading psychology.
No one believes that kids are robots but I will show in this post that they are.
Context For The Kids Are Robots Claim
Lewis [Lewis(1995)] accepts what might be termed the `standard model’ of emotions. That model holds that there are six basic primary emotions. These are the familiar emotions such as fear or sadness which are not complex in that they are not self-referential. There are in addition more complex emotions including shame which are held by Lewis to be self-referential, meaning that in order to experience that emotion, I must be able to introspect — to take myself as an object — which clearly `require[s] the concept of self’. Lewis holds that this ability to represent oneself to oneself becomes available `from around 18 months of age’.
Zahavi challenges Lewis’s claim that the complex emotions are necessarily self-referential. However, the challenge fails, as I will now outline. He notes that Lewis holds that mental states only become conscious when they are objects of introspection. Zahavi asserts that this commits Lewis to the `absurd’ consequence that animals and infants lack phenomenal experience.
But we can question Zahavi’s assertion. Lewis needs that further claim if the assumption is that phenomenal experience requires mental states to be available. They would need to be available to be taken as objects of introspection. That assumption is not indefensible, but it certainly needs defending.
However, an even stronger response is available to Lewis: he need not attack that assumption. He could instead allow it. Then he could argue that the consequence that animals and infants lack phenomenal experience is not absurd. He could accept that kids are robots. I will support this claim with two arguments, from anthropomorphism and evolution.
It is well-known that humans are remarkably prone to anthropomorphism. We tend to explain the behavior of even inanimate objects by projecting on to them emotions, knowledge, intentional states. In short, we project all the precursors of phenomenal experience.
So the refusal to ascribe phenomenal experience to animals and infants is not absurd; it is on the contrary wise. People see nothing wrong in making remarks about a chess-playing computer such as “it sees the threat and now it wants to castle.” They would go further and refer to the computer as `she’, if it had a female name.
There is a science fiction short story bearing this out. I will outline the story next. However, since in fact nowadays we have robot vacuum cleaners, we don’t need the story. I can confirm that everything in the story is what happens if you buy a robot vacuum cleaner. We call ours Reggie.
A Sci-Fi Illustration Showing That Kids Are Robots
The story concerns an observer, an engineer and a device he has built. The engineer has built a fairly simple small robot. It might look something like a vacuum cleaner of the type that rolls along the floor. The engineer has given this device some characteristics. It emits a series of somewhat anxious beeps as it scurries around, looking for a power source. Once it finds one, it extends a proboscis into the socket and contentedly hums as it draws power. Once sated, its lights brighten, the beeping ceases and the device moves off about its business, more quickly than before.
At this point, the engineer invites the observer to smash the device with a hammer. We and the observer greet this suggestion with horror. This is because we are convinced that the device has phenomenal experience. It is not adequate to object that this reaction occurs because we are reluctant to destroy property. The engineer is the one making the request.
The engineer could show us other examples of the device. He could demonstrate its construction and make it clear in all ways that it is just a machine. None of that changes our hardwired animist response. We consider almost everything in the world to be like us and therefore to be protected.
These tendencies are all the more liable to become engaged with animals (`Fido understands everything you say!’), and infants. Your reaction was perhaps short of horror. Perhaps it was reluctance. Even then, your options are to accept my claim or produce an explanation which does not rely on respect for property.
The Ease Of Anthropomorphism
Consider the plethora of anthropomorphic elements to the description I gave of the story. Was it strange or jarring when I described the device’s behavior as `anxious’ or `contented’ or was it entirely natural? This illusion is as potent as the Muller-Lyer illusion and works in the same way. We know that X is the case yet we perceive that not-X.
The conclusion in both cases must be to rely on what we know rather than what we perceive. We should require extraordinary evidence that any entities have phenomenal experience in view of our well-known promiscuous habits of painting it on to the world.
Finally, Nagel [Nagel(1974)] has argued convincingly that we cannot know what it is like to be another creature because we could not even aim for that target. The aim of imagining what it would be like to be a bat is approached by imagining ourselves with some or all of our characteristics and modes of perception removed and some or all of the corresponding items for bats added.
This is simply not the right target, which remains forever closed to us. The target is past an impenetrable barrier in cognitive space. It is no different to the one which prevents us from imagining life as a thermostat.
This argument of course does not show that bats do not have phenomenal experience. But it does show that we could not know if they did, thus greatly reducing the import of an argument relying on it being absurd that they do not.
If kids are robots, we would still think they are people.
All animals are subject to evolutionary pressure and experience extreme competitive stress in terms of energy budgets. This is true in terms of both physical and mental characteristics. Kaplan notes that
`individuals must live within finite energy budgets […] never spending more than they have available’.
Allocation of a finite budget entails trade-offs and hence forces decisions about the relative value of possible ways to spend.’ [Buss(2005), pp. 68-95] This budget must be expended also for mental characteristics:
`psychological adaptations are some of what humans have been selected to invest in, at an expense’.
[Buss(2005), p. 69]
Not only that, but the brain uses a lot of resource in mammals in general and humans in particular. In fact, the amount of energy a mammal obtains directly controls the size of brain it can “afford.” One citation on that is due to Hofman:
`adult brain size of mammals is a function of two major components: the animal’s rate of energy consumption and the evolutionary level of brain development’.
[Hofman(1983), pp. 495-512]
A larger brain is more complex, more expensive and more capable of providing or supporting more complex experience. That could include phenomenal experience.
No evolved individual uses energy and resources unnecessarily, where `unnecessary’ means in a way not promoting fitness. It is much less resource-intensive to simulate phenomenal experience than to have it. Actually having it achieves nothing. Simulating it produces immense benefits in terms of fitness. Human infants are not viable alone and require the support of adults. They can do this by simulation of simple phenomenal experience. And that simulation can be done by very straightforward heuristics. If kids are robots, they could still do it.
Crying Doesn’t Deny That Kids Are Robots
When hungry, it is important for infants that they make a noise which leads an adult to supply food. It is not important that there is `something it is like’ for them to feel hungry. Simple heuristics explain actual behavior and apparent phenomenal experience in infants and animals. Those denying that would have to attribute phenomenal experience to spiders. And they would have to say that sugar-eating bacteria `want’ to climb the sugar gradient.
Or they would need to conduct a difficult line-drawing exercise. They would need to discriminate similar organisms from each other. That is: two organism which have minor differences in cognitive abilities and yet major shifts in phenomenal experience. We may assume that the latter is a binary capacity.
Phenomenal Experience Is Expensive
One may object here by asking why, if phenomenal experience is so expensive, adult humans have it. This is of course too large a question to be addressed here. The topic has been widely considered, with questions ranging from `[h]ow could a physical system such as a brain also be an experiencer’ [Chalmers(1997)] to `what good is consciousness?’. [Dretske(1997)]
I will offer two observations, One: if we have it, it must be useful. Two: phenomenal experience could add fitness benefits. Those benefits could go beyond possession of correct information. Perhaps phenomenal experience makes it more likely that we will act on the information.
It is not an objection to i). to say that it would also be true of infants. In fact it is precisely my point that this is not the case. Infants do not need to do anything apart from make a noise when hungry.
Adults are not so simple. Dretske asks why we have phenomenal experience in relation to observation of sexually available members of the opposite sex. The key fitness benefit would be derived from the mere knowledge that they were so. If there is something it is like to know that, viz. pleasant and stimulating, then we could be more highly motivated to pursue the opportunity. Analogs of that argument may be run across all pleasurable activities. The same goes for unpleasant experiences.
Remaining with infants, what is needed to achieve their objectives when they cry? That it be unpleasant enough for the adults in earshot that they respond rapidly. What is not needed? Phenomenal experience in the infants.
So there are compelling reasons why we should choose the simplest explanation. No phenomenal experience exists in infants and animals.
We avoid the tendency to ascribe phenomenal experience widely and wrongly. Also, we avoid the claim that infants have a useless and highly expensive capacity. Thus we also avoid being on the wrong side of the theory of evolution. Finally, we may also note that none of us have convincing memories of undergoing phenomenal experience as infants.
Omission Bias is the tendency to think it is better to do nothing than to do something and make a mistake. We all have this tendency because it is a fundamental part of our psychology. It can be particularly strong when thinking about how much we will regret something in future if we make a mistake. We think we will regret doing nothing less than we will regret doing something.
This matters in financial markets, because it is not necessarily better to do nothing than to do something. The tendency to inaction creates inertia. We hold on to stocks longer than we should. We fail to buy new stocks when we should also do that. Formally, this is because is no practical difference between the following two scenarios. I could invest $100 in a stock that declines 10% the next day. I could also fail to buy one stock that appreciates 10% the next day. In both scenarios, I have lost $10. But Omission Bias will make me think that scenario two is better than scenario one. This is because in scenario two I have done nothing while in scenario one I have done something.
In this way, our psychology makes us prefer to do nothing. But the only important metric to judge the quality of my trading is the financial result. And that was the same in both scenarios: I do not have $10 that I would have had.
Adverse Effects of Omission Bias
A major adverse effect of Omission Bias is that it impairs your ability to assess performance. You will only look at what you did as opposed to what you did not do. This is because you are only worried about things you did that did not work. You don’t worry about things you didn’t do that would have worked because Omission Bias takes out the regret in that second case.
But the only measure of importance is how much money you made as compared to how much you could have made. So there should be equal attention paid to opportunities missed as there is to opportunities taken which did not work out.
This adverse effect is of crucial importance. Many investors do not have clear enough data of what has worked for them and what has not. It is essential to have a good focus on this for a number of reasons.
One benefit is that you can only manage your portfolio appropriately if you have been examining its performance precisely. A second benefit is that you might be able to identify some specific sorts of trade that you are particularly good at. You can then seek to identify relevantly similar situations and exploit them. Also — you might have a chance of avoiding disasters from the past occurring again!
Our ability to look at our failures and learn from them is also impeded by our natural distaste for thinking about the unpleasant — but failures are always more instructive than successes. One might almost say that any fool can succeed — but only an expert can fail well…
A major practical impediment to any attempts to correct for Omission Bias is due to the sheer scale of the problem. The number of shares you did not buy yesterday is absolutely huge. There is no way you can think about all of those. Nor should you. The more useful comparison is to think about the shares you could have bought or the ones you almost did buy. So that tells us that you should be looking at several buy options at a time. Look at what factors led you to choose the one you did choose.
Maybe you were looking at three oil companies. You compared them on price/earnings ratios, dividends and price/book value. You made a choice. Did that work out? (Don’t do this next day. Wait for a reasonable period. Otherwise you will just be looking at noise.)
Omission Bias is a sort of Agency Effect
What fundamentally is going on with Omission Bias is a sort of agency effect. If something bad happens and you could have prevented it but did not, this is seen as morally less culpable than if you did something which caused a bad outcome. After all, “you didn’t do anything.” I think this perception might be strengthened by the fact that the law says a lot about what we cannot do but rarely says anything about what you must do. You are at liberty to walk past a baby drowning in a pond. You are not at liberty to throw a baby in a pond.
This might be fine morally. But stock markets are not outlets for moral action. They are locations where you can profit. Or not. Bear in mind the possibilities of Omission Bias affecting your judgements of your own decision-making and your decisions will get better and more profitable.
Everyone wants to know what drives stock performance. It seems intuitive that over the long-term, stocks should correlate with GDP — but some have questioned this. In this article, I will aim to show that if you look over a long enough time period and don’t cut the data, that stocks are indeed correlated with GDP.
The first issue when seeking to examine the correlation between two variables is determining the optimal dataset to examine. This is highly contentious since it is quite likely that spurious correlations can appear in datasets that have been excessively manipulated either by cherry picking the date range or by eliminating outliers. For that reason, I will select the largest reasonable dataset and do nothing to it.
I think the earliest reasonable starting point is somewhere around 1950. Much before then, and one is looking at periods where the economy is so different to that of the current day, that it would be inappropriate to try to draw any conclusions. In addition, it is not that helpful to produce a model going further back and have to say, when asked about a prediction, that “this is what happens to stock performance when Hitler invades Poland…”
Here is a plot showing US GDP since 1950:
This immediately looks like some kind of power law growth with a kink for the global financial crisis. (We don’t really have enough data to look at COVID properly yet, but I expect it will produce another kink and not really disturb the trend line.)
This second plot illustrates stocks performance — it is the S&P 500 index over the same timescale.
This also shows some sort of power law rise, though it is clearly much noisier than the GDP curve. This does not however obscure the direction of travel.
Here’s a brief aside about polynomials. This is skippable if you don’t care what curves I am going to fit to the above two lines. Polynomials are curves of the form y = ax^3 + bx^2 + cx + d. The order of the polynomial is how many of the a, b, c, d coefficients are non-zero. So a first order polynomial is just a straight line of the usual form y = mx +c. The reason this matters is that we do not want to overfit or underfit.
If we underfit, we fail to capture information in the curve we are modelling. If we overfit, we get all the information in the curve, but we may add some features that are not really there. The best way to check for that is to see what happens if you extrapolate the curves beyond your training data. If you have overfitted, it will often be the case that your predictions go insane as soon as you are off-piste — i.e. outside the training dataset.
Here is an illustration of what under and overfitting might look like.
The blue circles are some arbitrary test data we want to fit. The other curves are the various orders of polynomial returned by a fitting package. As you can see, the first order polynomial — just a straight line as mentioned above — is totally inadequate and is a radical underfit of the data. The second order is not bad, but the third order is very good.
Fitting a Polynomial to US GDP
Fitting a third order polynomial to US GDP gives me the following plot.
I have slightly adjusted the data to make it fittable. The x-axis is now the number of days since 03-Jan-1950. I also zeroed out the US GDP for 1950 which I don’t think matters too much. You can see that the curve is a good fit to the data apart from a rogue rise at low day counts which again I don’t think matters because we aren’t going to try to predict GDP in 1950 — we already know it.
If you want to know the equation of that curve, it is :-5.265e-11 x + 4.351e-05 x – 0.2643 x + 414.1
If you think that the very small cubic coefficient means I don’t need third order, you are basically right but again I don’t think it matters as long as the curve extrapolates reasonably.
Fitting a Polynomial to Stock Performance
Let’s return to stock performance. Conducting similar manipulations on the S&P500 (recalibrating to days since the start of 1950 and setting the initial value to zero) gives me a different fitted curve, as below.
Note that in both cases, there are around 25000 days between the start of 1950 and today. (70 * 365 = 25,550).
Here, the equation of the fitted line is: 2.726e-10 x – 3.87e-06 x + 0.02366 x – 10.68
Again, the cubic coefficient is very small. The notable point about this curve is that it is very noisy. But it still follows a clear trend line.
Be Careful With Extrapolations
Now we get to the dangerous part. To look at stock performance long-term, we need to see if these curves behave once we extrapolate them. Let’s look at what they do if we double the day count to 50,000. (This is equivalent to making the date range double — so we are currently 70 years on from 1950; 70 further years added on takes us to 2090.)
There are two reasons this is dangerous. As I said, if the curves blow up, we have achieved nothing. The other issue is that you can’t extrapolate power laws forever. I will now discuss a brief example of that not working.
You will recall that in the early stages of COVID, people were plotting case counts and fitting exponentials to them. That looked like it was panning out to begin with. But the plot below shows what happens if you fit an exponential to the Florida case count as of yesterday.*
COVID as an Example of a Failed Extrapolation
You will recall that in the early stages of COVID, people were plotting case counts and fitting exponentials to them. That looked like it was panning out to begin with. But the plot below shows what happens if you fit an exponential to the Florida case count as of yesterday.*
It can be seen that the case count profile is in no way fitted by the exponential. So is it the case that the US economy can continue growing on an exponential basis indefinitely? Probably not. When will it stop doing so? No-one knows, but it might be beyond your investing lifetime. Also, bear in mind that it will have looked like it couldn’t continue growing exponentially at most points since 1950. We don’t know what technology or quantum computing or unimaginable developments are going to do.
I used the data science programming language Python to generate these plots. You can find put more about how that works, with a specific application to COVID, at the the link below:
So what happens if we extrapolate the curves we fitted? This is shown in the plot below.
This shows firstly and importantly that the curves continue to behave out to 50,000 days.
Conclusions on Stock Performance
Firstly, we can conclude that if the S&P500 and US GDP continue to develop as they have done, then by the year 2090, US GDP will have reached USD $90tn and the S&P 500 will have reached 26,000. That would be some great stock performance!
Secondly, I have not shown what happens if you try to fit curves to only more recent data. They are very noisy and that explains why a lot of analysis does not show any correlation.
Thirdly, this does not really backup Portnoy when he says “stocks always go up.” Partly that is the case because of the huge noise in the S&P 500 curve. Partly it is the case because you might have to wait a long time. Partly it is the case because the curve only shows that the S&P 500 always goes up. But the trend is your friend here if you wait long enough.
There are various conclusions that I have not argued for. I have said nothing about other countries. The above analysis would definitely not work for the FTSE-100 because that does not grow exponentially. It seems rather to exhibit a large saw tooth oscillation between 4000 and 7000 with a period of about a decade. That won’t correlate with anything. Similar points apply in Japan.
We have just far spent £337bn on the COVID response in the UK. This is reported by the FT at: https://www.ft.com/content/f0c7ab6d-33ba-4777-8dd8-0960a78a556a The virus itself generated much of this spend, but much of it was generated merely by the lockdown. The results of the Imperial model were a primary motivation for lockdown. That model was a Monte Carlo Simulation. I explain here briefly what a Monte Carlo simulation is and bring out one objection (among many) to the Imperial approach.
What is Monte Carlo Simulation?
Monte Carlo simulations exist to address a class of problems which do not have an analytical answer. Imagine I am in the pub and my home is 40 paces away. If I walk at two paces a second, I will arrive home in 20s. That’s an analytical question which has an exact answer. Here is a non-analytical question. I have drunk a bottle of tequila in the pub. The probability that I take a pace forward is only 50%; there is an equal probability that I take a pace backward. This does not appear to be analytical. You can’t say immediately what the probability is that I get home.
This is where Monte Carlo Simulation comes in. What you can do is simulate my journey home from the pub and see if I get home. (It’s called Monte Carlo because the method is based on random numbers, as we will see next.)
Sample Python Code
Here’s a really simple Python script called Randwalk that simulates the walk home. It’s called that because this is a random walk simulation. This sort of thing might be familiar to you from Brownian motion.
You can see that all it does is roll a dice 100 times and check to see if the dice shows three or less. That’s the 50/50 part. If the dice does show three or less, I take a step back. If the dice shows more than three, I take a step forward. This is repeated 1000 times, meaning I take 1000 steps in this version.
This entire code consists of a conditional statement in a loop. It’s extremely simple.
Output from Simple Python
We can then plot the output and we will see something like the below.
As you can see, a jagged pattern is generated. On this occasion, plenty of my steps were unfortunately in the wrong direction, and I never got home. But that won’t happen every time. I was luckier in the run below, or maybe I drank less tequila.
As you can see, here I was more than 40 paces north in this simulation. So I got home. (I haven’t bothered to clean up the code to stop running when I “arrive home” in order to keep it simple, but that could easily be done.)
Using Monte Carlo on the Random Walk Problem
So now we see how we can use Monte Carlo Simulation to answer the original question. What we need to do is run scenarios like the above a large number of times and see how often I get home.
Here is some only slightly more complicated Python called SimpMC which does that.
This just puts the whole of the previous code in another loop. So we do the simulation multiple times — that variable is itot = 10 in the code above. We then calculate the fraction of scenarios in which I get home.
Monte Carlo Results
This generates an answer of 0.2. But it is different every time I run it. Sometimes I get 0.3 and sometimes 0.4. That is happening because I have inadequate statistics. So let’s set the run number to 100.
Now I get: 0.14, 0.17, 0.21, 0.19, 0.15. Better but still not stable. Let’s set the run number to 1000.
Now I get: 0.195, 0.191, 0.208, 0.192, 0.205. That’s starting to get there. I am clearly converging on a probability estimate here. If I ran overnight, I would get a good answer.
Why is this an Objection to the Imperial Model
Finally to the objection to the Imperial model. Their code was unstable on multiple cores. Their response to this was “it’s a stochastic model so it will give different answers each time.” That response does not fly, as I will now explain.
Saying it is a stochastic model just means it uses this random number Monte Carlo approach. However — that does not mean it should produce different outcomes when run on multiple cores. It should not be unstable at all. The reported instability in the Imperial model is 80,000 deaths. This means that merely the error bar in the Imperial result is larger than the current total number of COVID deaths! — and that should not happen. To claim otherwise is to mix up the randomnesses. (I just made that word up but that seems fine.)
For sure, we saw randomness in the randwalk code — but that was just one run. When we did lots of runs in the SimpMC code, we started t0 converge. We got the same result every time in other words when we did enough runs. The Imperial model produces different results each time you run a large number of scenarios through it with the same parameters. That is equivalent to me getting different answers on the 1000 run version of SimpMC. If that happens, it doesn’t mean I wrote a stochastic model. It means I wrote a buggy model. Imperial have potentially cost us a lot of money here.
This article outlines the analysis of wine marketing. I will use as an example product Casillero del Diablo Cabernet Sauvignon. This is from the Central Valley, Chile 2018. Concha Y Toro is the producer.
This is a value wine which aims at a large volume of sales. The production is 22m cases a year. Concha Y Toro are among the ten highest volume producers globally. Cabernet Sauvignon is a very well-known variety. Chile is a very well-known country of production though maybe still slightly exotic and interesting for some consumers.
This wine costs £8 our bottle in Tesco. It is available online at £7.49 plus shipping. The wine is also available at Majestic for £6.99. That suggests that a special deal must have been done. The wine is positioned towards the top of the value end of the price range
The average price per bottle paid in the UK has been increasing slightly. It is a price-sensitive market, but there have been some signs of a trend to premiumisation. In previous decades, supermarkets would move a lot of wine at £5 per bottle, but Millennials are more health conscious. It seems inconsistent to them to go to the gym most days but also consume alcohol every day.
Many older customers would have drunk wine every evening. These customers would not pay £8. Millennials are replacing these consumers. Millennials are more likely to drink only at weekends. They are prepared to trade up.
The price of £8 in Tesco is perfectly positioned. It will hit the bulk of the current market. A Millennial professional picking up a bottle on the way home from work to drink with a meal would be a typical customer. They want to have something which is good value but with some interest to it to go with a meal or prior to going out.
The low-involvement consumer is the primary target demographic. This customer is slightly above the absolute minimum price-sensitive consumer but not willing to pay for additional structure or complexity.
Millennial consumers are likely to be reached by the marketing of this wine. It could sell well on the same basis to similar people in the US and more widely, though probably not in wine producing countries like Spain or France. Here, there will be too much local competition which will be offering fair quality without needing to carry a transport spend and marketing budget.
Wine Marketing: The Place
Ssupermarkets, convenience stores and online are the normal sales channels for a value wine like this.
Deep discounters will not carry this wine normally. They might do so if a special deal is available. This is because deep discounters will not want to pay for the heavy marketing/ad spend of a major brand like this. It is possible that deals can happen in some years though because 2.2m cases is a huge quantity. An economic downswing could cause the producer to be holding large quantities which will not really benefit from bottle ageing. In fact, this wine will likely deteriorate quickly (some commentators recommended drinking the 2018 in 2018 and no later in fact).
In this case, the wine is available in specialist wine retail though that is slightly surprising. It is possible that Majestic have specifically chosen this wine to capture the value-seeking customer. Or they carry it as a response to specific market conditions now. They could be moving very high volumes online to a lower-involvement customer base than would normally be the case.
Majestic could also be using a very good value wine as a hook to lure customers to their online space. They may then well be able to persuade them to trade up, but even if not, a sale is a sale.
Wine Marketing: The Promotion
Amazon describe the wine as the “UK’s No.1 Cabernet Sauvignon.” This is sticking strictly to the facts and is a positive message. People prefer choices already made by others since this seems to reduce risk.
Concha Y Toro make the following remarks on the product.
ORIGIN: Central Valley
SOIL: Mainly alluvial
AGEING: Aged in American oak barrels.
COLOUR: Deep, intense ruby red.
AROMA: The expression of cassis in this wine perfectly represents the Valle Central and its richness in fruit such as cherries and plums. The barrel ageing length thanks to the toast and coffee notes (sic)
PALATE: Medium bodied with silky tannins and long, ripe fruit and berry aftertaste, with impressive balance of fruit and polished tannins.
FOOD PAIRING: Red meats, well-seasoned dishes, and aged cheeses such as Gruyere or blue
There is a deliberate gradient of complexity to this description. The low-involvement consumer receives some information (cassis,richness, cherries) immediately. Low-involvement consumers could be switched off by the mention of tannins.
It is slightly confusing that there seem to be some issues with the English. One would not normally expect this in the wine marketing of a major conglomerate.
The “Story” in Wine Marketing
One of the keys to modern wine marketing is having a good “story.” Here, this means that people can wonder why there is a devil on the bottle. They will be intrigued and look into what the name means: why is it called “The Devil’s Cellar.”
Ladder branding is central to appropriate wine marketing. Yalumba’s wine marketing of Viognier is a good illustration of successful ladder branding. I will explore that in this article.
Yalumba is a successful winery which has occupied the Australian Viognier space. They have done this in a convincing fashion since around 1970.There are strong sales into the UK.The flagship super-premium member is “The Virgilius.”This is a very serious wine. The other members of the ladder gain brand equity from it. One criticism however might be that the ladder can seem to contain quite a lot of members.This could be a symptom of a tendency to overexploit a successful approach to the market.
Ladder Branding: the Members
Entry level — affordable
£11 — £15
Affordable / Stretch
Eden Valley Viognier
Yalumba Virgilius Viognier
£30 — £35
Flagship ladder brand member
Different price levels of Yalumba Viognier wines
The entry-level member Y Series is still not cheap. Certainly, Yalumba wish to avoid the opposite of a ladder brand benefit occurring where a super-premium member would not be credible if the entry-level product shipped at £5.
In addition, Yalumba have included a member (Organic) which could act as a bridge between the entry level Y Series and the pricier Eden Valley.
The ladder brand is co-marketed. Other products feature with it. This runs the risk of reducing the ladder branding effect. The Botrytis Viognier is really a separate product. Technically it is more expensive often than the Virgilius but this is a niche entry with limited supply available only in 375ml format.
There is also a Shiraz/Viognier product range which has only two members. They are also clearly located at distinct price points. Y Series is located at entry level and Hand Picked is located at premium level.
Also, Yalumba offer an Eau de Vie named V de Vie. This product uses Viognier grapes. V de Vie is co-marketed with the ladder brand even though this is really only coincidentally made from Viognier. There will be few varietal characteristics visible in a spirit.
The straight Viognier range does not really include these other products. However, they all go towards giving Yalumba a strong mind share of the Australian Viognier space.
The Eden Valley product is located just south of the premium slot. The name is well-chosen because it emphasises a sense of place but also there are positive associations between “Eden” and quality.
The Virgilius is available in Magnum. It is suitable for ageing. It is one of few Australian wines to be given a flagship location in the main wine displays at 67 Pall Mall, for example. So it has clearly established its premium nature.
Yalumba is not a listed company. It is therefore not required to provide extensive public data on revenue generation.
However, Wine Australia provide some statistics on Australian wine exports generally. Australia’s top five export markets by value as at year ending March 2020 are:
Mainland China (40 per cent of total export value)
United States of America (14 per cent)
United Kingdom (12 per cent)
Canada (6 per cent), and
Hong Kong (4 per cent)
Clearly China is of extreme importance in general. Nevertheless, there may well be an opportunity for Yalumba to move more Viognier there. Wine Searcher lists only one stockist in China. Eden Valley costs £75. This suggests there is effectively no supply and no information available to consumers in China.
The various rungs of the ladder brand are aimed at different consumers.
Above all, the Y Series is appropriately placed at entry level. It is interesting, fresh and approachable with decent levels of complexity without excess complexity. Most importantly, there is no sticker shock.
Consumers who are new to the brand will find nothing to put them off. So they may trade up to the upscale members on an appropriate occasion.
Many novice wine drinkers have moved from casual to serious interest in wine as a result of noticing the unassuming yet defined qualities of Y Series.
Placing a member (Organic) just north of entry level but still reasonably priced enough to be a potential step up is intelligent marketing. The customer is therefore more likely to step up the ladder. This is especially likely perhaps when a consumer has had the entry level product a few times or it is unavailable.
Making this wine organic and being very clear about that feature is also wise since Millennial Treaters will be a key segment here. That’s because the term “Organic” features prominently on their wish-list for products.
The pricing overlaps with the entry level product which again is intelligent because it encourages switching. The entry level product is sometimes available at a discount. Great care should be taken here. One should avoid the discounting of the Organic product so as to maintain its identity as “just slightly premium in a premium range.”
Eden Valley aims at consumers celebrating a special occasion. It is the “stretch” entry within the ladder branding. So customers who would normally stay on the lower rung might be tempted.
The Virgilius is the “aspiration” product within the ladder branding. Serious wine drinkers are the target. It is attractively priced in comparison to Meursault, for example.
The marketing materials include an immediate link to the wine club.
Moreover, there is an effective strapline. “Thoroughly captivated by this elusive, luscious and complex white variety, a visit to Viognier’s spiritual home saw the beginnings of a journey that would define Yalumba’s white winemaking future.”
This provides a strong brand story showing the roots of the product in the past. And the strapline also makes strong and clear statements about the qualities of the varietal. The consumer knows what to expect. The streamline also bootstraps from Old World experience with the variety. Consumers see that the product is made with care and experience.
In sum, we may conclude that Yalumba have very successfully employed ladder branding within their range of Viognier wines.
Wine SWOT analysis proceeds in the same way as in any other sector. This article presents a SWOT analysis for the particular case of Château Musar in Lebanon. See https://chateaumusar.com
Very strong brand story with significant historical underpinning since foundation in 1930s
Good level of brand equity especially among high-involvement consumers
Long history of organic production
Effective cellar door operation which attracts high-spending customers
Great ageing potential
Production can be described as “Bordeaux blend with a twist.” This is a strong combination of an established and well-known pattern but with some extra interest
Some competition but no other serious players in Lebanon currently
Good and reliable sales into France because of historical relationship between the countries
Extremely high quality consistent product
Good local supplies of skilled and unskilled labour
Distance of winery from vineyards causes risk of oxidation damage and heat stress to grapes in transit to winery
Whites are not close in importance to the flagship red blends and there is little serious portfolio diversification
Cellar door sales are somewhat impeded by the distance of Lebanon from western Europe and the US
Online operation exists but does not ship huge volumes for logistical reasons
Large volumes of water needed: not easily sourced
Can reach new consumers who are interested in exploring wines from novel locations. This is a major segment of the high-involvement market.
China has an increasing population in the mid-wealth range. There is though little penetration of premium wine from regions beyond those that are extremely famous such as Bordeaux or Burgundy
Explore movement of product into China via partnerships with apps which are very important in that market
Find ways of improving the logistical bottlenecks to move more product via online channels
Continued political instability in Lebanon
Peg of local currency to USD has broken meaning import costs will increase dramatically
Because of this, increased taxation is likely to result
Intensification of local regional conflict could impede winery and vineyard operations including labour supply and transport of product
Local competition does exist and could gain importance, diluting the position of Musar as the sole well-known producer in Lebanon; moreover if such competitors move product downmarket, this could dilute the soft brand equity of Lebanon
Poor internet bandwidth and consistency of power supply
Conclusions of Wine SWOT Analysis
The overall position revealed by a wine SWOT analysis is clear. Musar is in a very strong position provided the political/security situation remains stable.
This article explores wine online innovation by looking at the successful example of Gusbourne, an English wine producer.
Wine Online Innovation: Wine Clubs
Why might a producer choose to use a wine club and online sales to boost sales? What are the benefits and restrictions of selling wines this way?
A well-run wine club combines several major benefits for producers.
It creates a sense of exclusivity and buzz around flagship product which can optimise price achieved of the flagship product and sales volume throughout the range. Screaming Eagle have had great success with this approach. Their wine club has a long waiting list and is highly prestigious.
The mailing list forms a way of building lasting relationships with customers. They will often introduce family and friends to the brand. A wine club is a variant of a company newsletter but with much higher levels of interest generated among consumers.
Enhancing Customer Relationships
Customer loyalty provides some resilience against natural fluctuation in productions levels.
Customers who have a long-standing relationship will not desert the producer if there are supply issues in particular years. These can be caused by a poor harvest, for example.
Similarly, production may be larger than usual. In these circumstances, the existence of established allocations to customers helps shift a significant slice of product. It can ideally also be leveraged into extra volume at an (ideally minor) discount if this is seen as optimal under the circumstances.
Certainly, the wine club can be a conduit for driving traffic to the cellar door, which is a further useful channel in itself. It also further strengthens client relationships and is a positive advertising vector.
Valuable marketing intel comes for free. The winery can learn about its customers. It will learn their socio-economic breakdown. Market segments to address further will become clear.Which products appeal to which people?
Having the wine club online is a relatively inexpensive way of reaching a very large number of potential clients on a global basis.
Given the current nature of the commercial environment, not having an online presence or worse having an amateur one makes a producer appear extremely backward-looking. We cannot expect a producer with no website to be on top of the latest trends such as organic certification/new varietals post-climate change etc.
Having an online ordering possibility is extremely valuable in the extreme COVID circumstances currently obtaining. Wine writers have commented that small producers with no online presence are struggling enormously under lockdown because they simply have no way to move product.
There are no major disadvantages to having an online presence. Setting up and maintaining a website involves costs. This requires specialist skills. It is very important that the website be reliable and easy-to-navigate.
This includes consideration of how the website will appear on a phone. This is a channel of ever-increasing importance and many website today simply fail to be usable on a phone.
Staff must invest time in maintaining the website and potentially also a social media operation. For example, customers will often use social media to comment and complain about the product. Complaints must be handled on a measured and timely basis.
The entire nature of social media posting is confusing for two reasons. Firstly, one must somehow combine professionalism and informality. Secondly one must stay abreast of an ever-changing landscape of platforms. For example, the median Facebook user in the US is now 41. This may be fine for reaching the established client base but will increasingly miss the future Millennial market base.
Keeping the online stock updated is another task which will involve sustained effort .
Wine Online Innovation at Gusbourne
Gusbourne is a great example producer that is using online innovation to sell their wines.
Gusbourne is a successful producer of English sparkling wines which has an online wine club named “Gusbourne Reserve.”
This is an intelligent choice of name which combines a sense of exclusivity with simplicity and a focus on the brand; it would have been very easy and lazy to call it the “Gusbourne Wine Club” etc
Consumer Touch Points
How does this online presence gives the producer various touch points with the consumer?
The navigation options at the top of the start page of the site reflect the touch points. These are as follows.
TOURS & TASTINGS
TIME WELL SPENT
50% of enquiries die when a question is asked or input is required on a website. It is therefore essential to have as little time/navigation required between arrival on a site and an opportunity to purchase as possible. Therefore, many of the options above lead on to immediate revenue-generation opportunities. The rest offer further information which will ideally retain visitors on the website.
The first four options offer further information. Gusbourne can thus appeal to potential clients who are specifically interested in the location of the vineyards. Clients want to know how the wine is made and what wines are available. The very first option ABOUT US being the default for someone who wants to know more but has no specific direction of enquiry at present.
All of the remaining options are revenue-generation opportunities. Details of where to buy the wine offline are provided. Opportunities to visit the cellar door are promoted. Further details of the wine club are offered. TIME WELL SPENT is a COVID opportunity whereby their sommelier will engage on specific topics of interest such as food matching via bespoke online channels such as Zoom.
The remaining two options BUY and TOURS are placed at the far right. These are repeats of options already available but are basically designed to catch very busy people. If someone only reads one word on this page, it is likely to be one of these two. One offers the chance to buy product and the other offers the chance to book a tour. Both are good revenue opportunities for Gusbourne.
Effective Wine Online Innovation
How effective is Gusbourne’s online strategy?
The online strategy of Gusbourne is extremely effective as a result of various factors. The site has strong design with varied professional photography. It has frequently updated content. The last two posts are from the same day and six days previously. Appealing photographs front both posts. There are nine posts from the current month which is impressive.
The content is fresh and informative. There are discussions on what is happening in the vineyard in spring with a theme of “hope” and renewal. This is particularly valuable and sensitive to current conditions.
There is plenty of food matching advice from chefs as to what can work with the product.
There is a post on the unconventional and interesting background of the founder. He was a surgeon from the noted wine region Stellenbosch in South Africa. This adds human interest which is important to building the brand story.
This is very strong offer with no wasted words. It tells potential members what they are getting and why they should want it. It is good marketing to explain to the customer why the product will give him what he wants and tell him also that he wants it just for the avoidance of doubt.
The website sells an “experience” or how a customer will become more the person he wants to be or to project. “I want to be the kind of person who is in the club.”
“Become a member today to guarantee your allocation of our wines, direct from the cellar, and access to a range of exclusive benefits.”
There is a strong use of consistent livery across the website. This enhances its professional appearance. The front page of the Reserved subsite has very strong and bold graphic design. This is simple and straightforward and emphasises the product.
Text is sparse. This is wise since “busy” websites are extremely off-putting. There are some navigation options and the mission statement below.
Allocation Appeal in Wine Online Innovation
Immediately below the main page of the subsite comes a USP statement as below.
“YOUR ALLOCATION: Two bottles each of Gusbourne Brut Reserve, Rosé and Blanc de Blancs are allocated to each member at cellar release and delivered in two cases of six during the course of the year.”
This tells clients exactly what they are getting but also makes it clear that the initial commitment need not be immense.
The offer continues as below.
“Throughout the year, you can order additional bottles at preferential rates, and you’ll also have the opportunity to order our limited edition, mature and rare wines, which are only available to members”
These exclusive benefits are likely to be extremely appealing to clients. Note also the constant address of the potential client as “you.” This emphasises what it is that YOU are going to get if you join Gusbourne Reserve.
“All orders receive complimentary UK mainland delivery”
This is a way of giving a discount to loyal customers without actually reducing the price of the product. Discounts can have adverse market implications.
Gusbourne is not cheap. This is because low yields in the UK under the current conditions mean the price per bottle will be relatively high at ca. £20 per bottle. The winery has to produce a premium product to be viable at this price.
Social Media Operation
There is a comprehensive social media operation on Facebook, Twitter and Instagram. These have dedicated content tailored to the different platforms. Twitter is less visual than the others; Instagram uses more video than still photography.
The mailing list is easy to join. No commitment to joining Gusbourne Reserve is required. The upgrade/conversion rate will be good given the effectiveness of the site design.
This is such a professional and widely-ranging online operation that it is difficult to suggest any further improvements.
One possibility to explore would be to add more video content. The current site is very text-based. This may gain less traction with younger people now and in the future.
Loss of Wine Production is a difficult problem for a winery. A wine producer may have many loyal customers. Normally, the producer sells all wine produced. Suddenly only a reduced volume of wine can be sold. This could be catastrophic weather destroying much of their grape crop, for example.
In Argentina, hail is a particular threat to winemakers:
In, fact loss of wine production causes $10bn of losses to winemakers every year. If wineries seek to maximise revenue and minimise losing loyal customers, what are some of the tactics they might apply?
It would generally be prudent to store some wine for two reasons. Firstly, the winery could handle production outages such as the above-mentioned circumstances. If any wine can benefit from ageing, then some should be stored in any case. That is because storage will maximise value, though potentially at the cost of some temporary cash-flow issues.
If however, the producer sells all the wine, the problem is more difficult. If the producer sells all the wine every year, then there will be no reserves available.
Using Loyalty as a Metric
“Loyal” customers may receive all the wine. If so, then there is no way of providing the desired amounts of wine to all customers who are loyal. If on the other hand any customers are casual, they should be the first to lose their allocation.
This suggests an approach whereby the winery should rank its customers by loyalty. Loyalty means that a customer has bought a large quantity of wine for many years.
A loyalty figure of merit is calculated by multiplying the number of cases bought and years for which they are purchased. However, the simplest way might be to find the total of cases each customer has bought in the last yen years.
We should also adjust this figure for recency. It would be a mistake to give the same loyalty score to one buyer who took 10,000 cases ten years ago and nothing since and to a second buyer who took 10,000 cases last year and says they will do the same in future. It is easy to do this. Divide the number of cases by the number of years since they were bought.
This algorithm produces a rank order of loyalty which can now feed in to allocation decisions.
Process for handling Loss of Wine Production
The first task is to contact all customers and explain the position in terms of lost production. The initial aim here is to find out if any customers have alternative supplies of a reasonably similar product
The winery should seek to proactively identify alternative suppliers. These could be located in any major global wine location. Those suppliers could have product/excess productavailable. If so, a deal is available.
The winery could open discussions with this other producer about supplying existing customers. However, a major risk here is that customers remain with the alternative supplier in future years. Therefore the winery should consider actually purchasing alternative supplies and thus retaining client contact.
Such an operation need funding. Banks may lend, but if the winery has crop insurance, this is a good application of such receipts.If it does not have crop insurance, it should acquire some unless it is in sufficient funds to self-insure against this risk.
The secondary aim is to “take the temperature” of clients. They may be unhappy about the problem. Or they may be philosophical about it. Agricultural production can not be guaranteed.
Loyal, understanding clients may defer their allocation for this year. This could be on the basis that they will receive bonus allocations next year if they so wish.
Alternative production methods
A specific shock could impair the grape harvest. Hail in the exact vineyards owned/operated by the winery might be the problem. If so, there may still be other grapes available in the region from other producers or specialist growers. Bought in grapes would allow the reinstatement of production. If the adverse conditions affected the entire region, this will not be a solution.
If the adverse factors are different, there are likely to be ways of addressing them, which will cost money. A temporary bottling line can replace a damaged one.
Loss of staff could be the issue. Alternative staff are available from different locations. The vineyard may be suitable for mechanised harvest.
If for example a global pandemic makes it impossible for workers to travel from traditional locations in CEE, then it might be possible to employ workers furloughed from other businesses. Hospitality is one obvious source.
The revenue maximisation question is secondary because revenue is to some extent already maximised. The winery sells all of the wine available.
Some clients may pay more in order to receive an increased allocation. This is risky since it could appear to be price gouging. Clients need transparency here. This should emphasis that the winery has certain fixed costs. It also has a much reduced production this year, so it needs to charge more per bottle in order to survive.
Clients amenable to this could receive discounts or guaranteed priority in future years.
The winery could enter into long term supply contracts with top clients. The clients are guaranteed what production there is in a particular year.
The FT today runs the argument that there are serious adverse effects of the lockdown which should be considered when deciding when to lift it. It could be that eventually lockdown kills more people than it saves.
These include the exacerbation of poor mental health and the way that people who are victims of domestic abuse have no escape.
Not a bad argument as far as it goes, but it is open to the objection that lives and cash are incommensurable. I wouldn’t make that argument, but many would. This objection points the way to a stronger argument in the vicinity.
Asking the Right Question about Whether Lockdown Kills
Lockdown will kill more people than it saves. The choice is not “accept some economic damage in order to save lives or not.” It is “kill group X or group Y.”
We kind of get there already with the estimate by the oncologist Sikora that excess cancer mortality caused by lack of screening will be 50,000. But beyond that, the economic damage caused by the lockdown will be so immense that it is likely that the second round effects on future NHS spending will also be huge. That will again cause many more deaths.
Moreover, those deaths are more damaging. We may think we are killing Group Y in order to give Group X a few extra months. Most of Group X are over 80 and with existing health conditions. Indeed, 50% of COVID mortality takes place in a care home setting, where the median length of stay is in any case 12 months. Group Y is younger people who would be economically active for a lifetime. If they are killed by the lockdown, they will not be. Note that a 1% increase in unemployment causes a 0.79% increase in suicide.
However, against this it should be noted that men who smoke and are aged 80 with a morbidity factor such as diabetes still have remaining life expectancy of 5 years. This rather extraordinary fact suggests that some men who have made it that far are very resilient.
My second reason for scepticism about the lockdown is that we are not really doing it (in the UK). There is massive non-compliance. Many people cannot be bothered to maintain a 2m separation even when it is easy to do so. There are signs on Battersea Bridge stating that one should cross on the leftward pavement. This is too difficult for about 20% of people. 25% of people admit that they are non-compliant, which makes me believe that in fact, maybe 75% of people are actually non-compliant.
Also note that subject 31 caused 81% of cases in S Korea. This implies you need a 100.0% compliance rate to do anything useful. That also makes me question the point of lockdown. Though of course I am complying with it because we cannot be sure whether it works.
The primary argument for lockdown derived from the Imperial model. This is a questionable piece of work. It is pure mathematical modelling, not a result of data analysis. The model is not experimental observational science. It is also “5000 lines of unaudited C from 13 years ago.”
Lockdown appears more reasonable based on the latest data, however. One powerful argument against lockdown was that India was clearly incapable of doing any kind of stringent and comprehensive lockdown. Indian mortality rates were mild however. That is no longer true. It may be the case that we have to do lockdown even though lockdown kills.