Author: Andrew Veitch (page 1 of 2)

Digital first?

There are an enormous number of marketing job adverts that say the prospective employee should have a “digital first” outlook.

This is something we disagree with quite strongly. The approach should be “cost per customer acquired first” which may well result in the choice of a digital channel but that is a completely different thing from starting with a prejudiced view that the digital channel will work best.

Nowadays the main digital advertising channels are based on Generalised Second Price auctions. In an ordinary auction (called an English auction) a product will be sold to the highest bidder at the price they bid. In a GSP auction, the first place will go to the highest bidder who will pay the price of the second bidder and so on. (Bids are adjusted by Google based on Click Thru Rates as of course the highest bidder may not receive clicks and will therefore not have to pay for their bid.)

The good news is that the GSP is sightly less prone to the Winner’s Curse than the English Auction (as we need more than one bidder to get their bid wrong). But it is still the case that the number of participants in the auction will tend to increase the price per customer acquired.

At Move Fresh we have found that many of the PPC auctions have now increased so much in cost that we are much better buying more conventional media (such as inserts).

The point is that we are not a digital first marketing company: we are an acquisition first company. We are completely agnostic on media. We think “digital first” is as dangerous as suggesting that a company should focus purely on traditional marketing.

It’s what works.

Shakespeare on Copywriting

I am very pleased to be spending much of my time now teaching copywriting and marketing to our very talented team.

How does one write copy for Google Adwords or for the other e-commerce channels we use?

Looking back over the course I’ve been teaching I’ve realised that the vast majority of it has been quotes from William Shakespeare, followed by William Blake.

There have been questions about why e-commerce marketing which started in 1995 has been almost entirely covered in my course by Shakespeare (1564) and Blake (1757). Am I not at least more than two centuries out of date?

The answer of course is that human beings haven’t changed that much. And indeed some of the course has even gone back to the Ancient Greek Rhetoric.

If you understand anaphora, anthimeria, archaism, assonance, asyndeton, brachylogy, chiasmus, diacope, epanalepsis, epimone, epistrophe, hyperbaton, hyperbole, irony, malapropism, metaphor, metonymy, onomatopoeia, paradox, paralepsis, polysyndeton, simile, syllepsis, synecdoche, tautology and zeugma as well as Shakespeare and Blake then you are well on your way to writing a great Google AdWord advert even if you are 200 hundred years after the masters.

Monty Python on Statistics

I was reminded today of my favourite Monty Python sketch on statistics. I had a look for it on YouTube but sadly couldn’t find it.

The sketch was set on election night with a reporter who did a vox pop and asked a lady in the street how she was going to vote. She said “Conservative”. They then went back to the studio and extrapolated this to their swing-o-meter which predicted a 100% swing to the Conservatives and all seats in the House of Commons switching to that party.

It was very funny. But like the best comedy it was also very true. We’ve all been in situations in business were very small datasets are extrapolated.

Interpolation is of course much more accurate. But to interpolate my experience of statistics in business I can say that interpolation is something that is much less common that extrapolation.

Michael Fish

Michael Fish announced on TV on 15 October 1987 that there was no hurricane on the way. That evening the worst storm to hit South East England caused record damage and killed 19 people.

Andy Haldane of the Bank of England recently said that the failure to predict the 2008-09 financial crash was a similar moment for economists and has resulted in economics forecasting being “in crisis”.

It is true that the record for prediction of economists, financial analysts and accountants is very poor. When we look back over the last half century it is indeed true that all three groups failed to predict the big crashes in US public companies.

But conversely they also failed to identify the ten days over the last fifty years which accounted for roughly half the return of the same US public companies.

In other words, out of 18,250 trading days just 10 days accounted for half of the return. So surely it was a failure of economists, financial analysts and accountants that they did not identify these days in advance? Or identify the days in 2008-09 that resulted in dramatic reductions in market value?

I would disagree with Andy Haldane on this. I don’t think it is reasonable to expect economists to identify in advance the 0.05% of days in which markets move dramatically.

My favourite book on this subject is A Treatise on Probability by John Maynard Keynes who wrote in Chapter 3:

Is our expectation of rain, when we start out for a walk, always more likely than not, or less likely than not, or as likely as not? I am prepared to argue that on some occasions none of these alternatives hold, and that it will be an arbitrary matter to decide for or against the umbrella. If the barometer is high, but the clouds are black, it is not always rational that one should prevail over the other in our minds, or even that we should balance them, though it will be rational to allow caprice to determine us and to waste no time on the debate.

Keynes understood that some probabilities are measurable. He gave an example in the book of pulling red and black balls out an ern of which half are red and half are black which is clearly a system that can be easily measured but many probabilities are inherently unmeasurable such as his example of the requirement for an umbrella or his discussion of an ern where the the proportions of red and black balls are unknown.

However although there are many probabilities that are not measurable but are comparable. The example Keynes gave was the possibility of surviving a walk home in a thunderstorm: this is not measurable but it is clearly less safe and therefore comparable to walking home during good weather.

So to get back to our trio of economists, financial analysts and accountants. It is unreasonable to expect them to expect any of them to predict a dramatic period of recession or growth however it is reasonable to expect them to suggest that a situation is more or less comparable i.e. more or less likely to result in a period of recession or growth.

I do not think that it is a failure of economics to predict the ten best trading days of the last half century. I do think that economics is useful in producing data that makes the best trading days comparable with poor trading days or the best investment opportunities comparable with the worst while accepting that neither are measurable.

So in my investments I accept that I cannot make predictions as to when it will rain. But I can see clouds in the sky as Micheal Fish did successfully over his career at the Met Office from 1962 to 2004.

Bertrand Russell on Investment

Back in my youth I did a course on philosophy and one of the books we read was Bertrand Russell’s The Problems of Philosophy. This is a quote from Chapter 6 – On Induction:

Domestic animals expect food when they see the person who feeds them. We know that all these rather crude expectations of uniformity are liable to be misleading. The man who has fed the chicken every day throughout its life at last wrings its neck instead, showing that more refined views as to the uniformity of nature would have been useful to the chicken.

Philosophical induction is of course a very different beast from mathematical induction. In our home town, Edinburgh University have created a programming language Haskell which can be proven correct by proof by mathematical induction. Mathematical induction is based on proofs rather than the slightly flimsy enumerative deduction in philosophy which is based on observation.

Going back to Russell he creates a severe problem for the Chartist. This is a very common investment mistake and I will admit that I have fallen for it myself in the past. It’s very easy to look at a graph for a stock and assume that this has some link to the future.

An individual example would be Enron which of course won many awards before collapsing to zero. A bigger example would be the pre-revolution Russian stockmarket which outperformed all of its competitors for a whole century before total wipeout. Both these examples are very negative but the failure of proof by induction also applies in the other direction. An obvious example would be Walmart which went for a few decades in a fairly flat way before completely taking off.

So next time you are looking at a graph and expect it to continue on its happy trajectory remember the story of Bertrand Russell’s chicken. The past is not a predictor of the future – good or bad.

5 Vital Brand Attributes

One of our advisors has five things he believes a brand has got to get right in order to succeed:

  1. Show expertise – In the old days the only space available to show expertise was on the label which was limited to put it mildly. But customers now expect content (on web, apps and in print) that shows that a brand knows what it’s talking about.
  2. Personalisation – The day of the mass market brand has gone. Consumers now expect something unique for them such as Tails, Moonpig or Diet Chef with our unique recommendations and 360,000 combinations available for the diet.
  3. Trust – This is particularly important in food where the industry has generally got quite a bad reputation. Innocent would be a great example of a brand who have worked for deep levels of trust. Trust is of course earned over years but can be lost overnight.
  4. Care – This is quite a deep one covering production of the product, impact on the environment and the community and of course care of the customer.
  5. Innovation – It would be fair to say that the food industry has not got innovation deep in its DNA. Even heritage brands needs to innovate while preserving their provenance.

Vertical Integration in Grocery E-commerce

This is quite a hot topic in the sector. In general I’m not a fan of vertical integration, I much prefer a tightly focused business. However I think there are two specific instances where vertical integration makes sense:

  1. Customer experience – Certain processes are vital to the customer experience and are a true differentiator such as the website, mobile apps and fulfilment. However backend systems such as accountancy and warehouse management should just be bought rather than owned. Historically processes that were differentiators have become commoditised so it may be that outsourced options will emerge for the current differentiators in the future (but no doubt other differentiators will emerge).
  2. Using unique technology – Nespresso have (literally) hundreds of patents on their technology and use a very complex process to create their capsules. This is not something that could be outsourced as they would be risking their intellectual property and in addition the only possible customer would be Nespresso which would prevent the manufacturer from sharing costs with multiple customers.
  3. Provenance – Guinness or Perrier would both be good examples. Indeed Guinness did experiment with brewing outside St James’s Gate, Dublin but that beer was considered inferior and production was shifted back. Equally Perrier could not be outsourced to another water company. It is a key part of the respective brands that these products are made in a particular place under the control of the company.

I think there are some quite bad reasons for vertical integration:

  1. Margin/cost – Unless there is a market failure margin will typically land at cost of capital. Margin in UK food manufacturing is very low and it is almost certainly not a good use of capital for an e-commerce business. Committing to larger volume is or outsourcing to a lower cost country is nearly always a better way of reducing price than vertical integration.
  2. Reliability of supply – This is quite often given as a reason. I would accept that in certain locations (e.g. Tata running their own electricity supply in India) this may be valid. But with careful sourcing reliability is rarely an issue in a developed country.

Move Fresh’s core competencies are management of brands, sales, marketing and creating a great customer experience. Outside of these areas we are keen to find high quality partners to fill in the gaps.

Deal & Dealmakers Finalist

We are delighted to announce that Move Fresh has been nominated for MBO/MBI of the Year at the Deal & Dealmakers Awards for our acquisition of Diet Chef.

Supporting Django

We divide our technology requirements in two:

  1. Backend systems such as finance or warehouse management which are commoditised. We buy these systems off the shelf.
  2. Customer facing systems such as mobile apps and our website which are key differentiators. These systems we develop ourselves.

Our own platform runs on Django which is one of the more modern frameworks (incidentally it is also what Ocado use). We are delighted to announce that we are supporting Django by sponsoring the Django: Under the Hood conference in Amsterdam, 3-4 November 2016. We do think it is important that commercial users of open source software put something back into the community.

Our E-Commerce Business Model

At Diet Chef we achieved a total return on equity of 8,999,900% over a three year period. If we had the same growth over the next three years we would be ten times larger than HSBC, the biggest company in the FTSE100.

Sadly, this is an outcome that will not happen. It’s just not possible to deploy the larger capital that we have today as efficiently as we did in the old days.

However it is worth looking at the business model we used as this is still very relevant.

First off, most of the funding for Diet Chef was not provided in the form of equity. Most of the funding actually came from suppliers who initially sold to us on 30 days while we received cash from customers within one day. This negative working capital requirement then scaled up as the business grew.

For our first TV campaign we requested that suppliers increase their terms to 60 days to fund it. They agreed to this and the result was that both our business and the suppliers benefited hugely.

This is quite a well worn path in retail, however in e-commerce there is less capital expenditure required to support growth (shops, etc) so the equity requirement is substantially reduced.

It is also a poor man’s version of Warren Buffett’s strategy of investing his insurance float, where the insurance premiums are invested over the years between receipt of the premiums and payment of the claims. Thus he makes a double profit: once on the insurance business and once on the investment of the premiums.

Back to Diet Chef, the other key was that we recruited customers so that they were profitable within 60 days. Generally, in direct businesses it will take a longer period for customers to become profitable which can result in a form of overtrading. This allowed us to grow the business to market capacity pretty quickly.

It’s not realistic in all e-commerce businesses to recruit profitably, particularly those in more mature markets. However these businesses should all have databases which will provide a profitable retention business which should then fund the loss making recruitment business.

There were of course a lot of non-financial reasons for the success of the business which I will cover later. But getting the business model right did allow us to scale the business up dramatically with no external equity funding.

Kevin and I have both spent a lot of time in Silicon Valley and one of the interesting things there is the focus on business models. But it’s something that is seldom discussed in Britain.

Older posts

© 2017 Move Fresh

Theme by Anders NorenUp ↑