Attention Span of Humans is More Than That of Goldfish

If you live in the marketing world I’m sure you’ve heard about the shrinking attention spans of human beings and that now the attention span of human beings is less than that of Goldfish.

The source of this data is a 2015 report by the Consumer Insights team at Microsoft Canada. The study was based on a survey of 2000 Canadians and also studied the brain activity of 112 people as they carried out various tasks. The statistic on the 8 second attention span did not come from the study. The report quotes its source as “Statistic Brain”. Statistic Brain source all their figures but the sources quoted don’t have any record of the statistic. Its literally a factoid which the dictionary defines as “unreliable information that is reported and repeated so often that it becomes accepted as a fact”.

Interestingly, in the world of neuro-psychology, Goldfish are considered a model system for studying process of learning and memory formation. They actually have a good memory!!!! Coming back to human beings, psychologists will tell you that attention spans are a function of the task at hand. The attention span engaging with billboards is different from the attention span writing this blog.

This unfortunately is not an isolated incident. At marketing conferences or in marketing literature, you may have heard that the human brain process visuals 60,000 times faster than text – again there is no scientific evidence for this.

As human beings we have a tendency to over generalize and this opens up an avenue for creating new business’s and brands. Take for instance the 10,000 steps a day regime which has fawned an industry of wearable devices. The 10,000 steps is a completely arbitrary number which has no scientific basis. To tap into the buzz around the Tokyo Olympics in 1964 a company called Yamasa designed the worlds first step counter device and called it manpo-kei, which translates to a “10,000 – step meter”. So the number 10,000 is nothing but the result of a successful Japanese marketing campaign in the mid 60’s!!! 

As market researchers we must resist the temptation to become instant believers in the latest headline without pausing to check the veracity of the source and the methodology. Lets take the latest headlines doing the rounds in the marketing and business press – “Nike sales increase 31% after Kapernick ad”. Lets unpeel this “growth”. 

• When did the growth happen - The articles start with a bunch of opinions on the ad (which I happen to love) and somewhere at the end will mention that the 31% increase is during the Labor day weekend (September 2^nd to September 4^th)  – really a 3 day window!!!   
• Where did the growth happen - The source of the data is Edison Trends which uses a panel to determine online sales. Online sales accounted for 6.5% of Nike sales in 2017. So we are talking of an increase on a very small proportion of total sales. 
• What is Nike’s ecommerce growth – Nike’s online sales grew by 29.7% in 2017 and is growing at 25% in 2018. The growth of 31% appears to be in the same ballpark.

So the growth is over a 3 day window, covering under 10% of the total sales and is in the same ball park as the growth over other weekends – the market researcher in me says we just don’t have enough data to attribute sales impact to the ad. I am sure Nike has information but its just not available in the public domain. And yet the marketing journals have concluded that there was a 31% increase in sales.

Coming back to the Title of the blog – there is no scientific evidence that this is true but if this blog goes viral then I would expect someone to headline this at a marketing conference!!!

Lessons from The French and The British Concurrence that Rescued the Market Research Industry

After a string of horrific polling related mistakes, the market research industry can bask in new found success after getting the French and British election right. For once the British and the French worked in concert and it helped the market research industry.

On the day after the US elections, I remember speaking to a colleague and saying “if only we had a dollar for everytime the opinion polls were wrong we would be laughing our way to the bank”. So on the day Theresa May called a Snap election, I decided to put money on what I thought would be the outcome of the election – a hung parliament. I put £50 with odds of 11:2, odds for a Tory victory was 1:7.

I told some of my colleagues about my bet and was told, in polite language, that I had lost my marbles. Having put my own money I followed the election campaigns quite closely. There are three lessons that that I see being relevant to the Market Research Industry

1. History repeats itself -

“All men make mistakes, but only wise men learn from their mistakes”. Winston Churchill

A quick look at history shows that snap elections have in only half the occasions resulted in a win for the party that called the election. A snap election, however, for a “mandate” has on both past occasions resulted in a hung parliament. Clearly, the British public expect the leaders to go ahead and do the job. They have little respect for politicians who “ask them for permission to do their jobs”. Why would 2017 be different – it wasn’t.

Year	Status	Reason for snap election	Outcome
1923	Tory	Mandate to raise tariffs	Hung parliament
1951	Labour	Increase majority	Tory victory
1955	Tory	Called by new leader post Churchill’s retirement	Tory victory
1966	Labour	Increase majority which was reduced to two	Labour victory
1974	Tory	Mandate to face down miners strike	Hung parliament
1974	Labour	Win a majority – Labour were a minority government	Labour victory

Learning for market research – It pays to look at history to identify learning's that could help in designing the study and more importantly explaining the results from the study. I am not saying that  you cant change the outcome but being aware of pitfalls and opportunities gives you a leg up. To draw an analogy – if you are trying to put a man on the moon it would be foolish to ignore gravity. Acknowledging gravity does not act against putting a man on the moon. If anything acknowledging its presence aids in achieving the desired outcome.

2. Experience does not guarantee expertise  –

“It is a mistake to speak of a bad choice in love, since as soon as a choice exists, it can only be bad”. Marcel Proust

I ignored the 18% point “lead” that Theresa May had on the day she called the election – why? One reason – Context. 

Context is the reason why more luxury cars are sold at yacht shows than at car shows. If you ask me if I would like a car that can reach a speed of 100mph in 1 second – the answer is yes of course. But that is the wrong question – since I live in London I rarely, if ever, reach more than 30mph and so the feature is irrelevant to me. Brexit was done and dusted, it was not relevant in the elections. People are now concerned about rising prices, the state of the NHS, trains etc - those are what they want politicians to solve.

Learning for market research - Market research is a specialised area where expertise can’t be automated and must not be ignored.

I have not seen anyone who is not a Doctor, looking at an MRI scan and offering a diagnosis or a prescription. But there are enough people who look at a statistic (18% poll lead) and offer a conclusion and advice. I often hear people say “numbers tell a story” but funnily I have never heard number 5 have a conversation with number 9!!!

In the market research industry, expertise is required to frame questions and interpret the results. I am not talking about knowing about the interpretation of interaction in contingency tables (also known as the Yule-Simpson effect) – it explains how you can gain share in every segment of a market but lose marketshare at the total market level (I’ll let you think that one through J). Of course, knowledge of the Yule Simpson effect is useful but not enough (or necessary) to qualify as an expert.

A prediction of the future on the basis of the past assumes ceteris paribus which is not a real life possibility.  The tack that Theresa May took i.e strong and stable vs chaos was identical to what David Cameron had taken two years back but the context had changed, hence the result changed.

  

You do need to have expertise to understand the context in which questions have been asked, read the numbers correctly and advice on the basis of that. The research industry needs to reposition itself as the consultants who speak for consumers. The larger question is whether it is a consumer insights industry or a market research industry.

3.  Forecasting works but is dangerous under uncertainty and for the long term –

“You can always trust the untrustworthy because you can always trust that they will be untrustworthy. It is the trustworthy you cant trust” Captain Jack Sparrow, Pirates of the Carribean

A year before the French elections Alain Jupe was headed for victory with Emmanuel Macron not even in the fray. How things change in a year !!!

Clearly the long range forecasts were wrong and a lot of it had to do with the volatility. By contrast, once the candidates were chosen for the runoffs, the forecasts were remarkably accurate and consistent over time – probably because of the lack of volatility.

Learning for Market Research  - In his seminal book , The General Theory of Employment, Interest and Money, Keynes said we need to “devote our intelligences to anticipating what average opinion expects the average opinion to be”. He was speaking about the stock markets but the same holds true in market research when dealing with stable conditions. When we are looking at the 20^th flavor of a well established brand of biscuits or the 15^th color of a sports shoe, its easy to make a prediction.

However, looking at the demand for VR in the future is tough. As a reminder (history teaches us lessons!!!) here are a few famous predictions gone awry

·        “We don’t like their sound and guitar music is on the way out”. Decca Records about The Beatles

·        “The world potential market for copying machines is 5000 at most”. IBM to Xerox in 1959

·        “There is no reason anyone would want a computer in their home”. Ken Olson, Co founder of Digital Equipment Corporation. 

If those sound prehistoric remember how Steve Balmer (CEO of Microsoft) wrote off iPhones “....is the most expensive phone in the world and it doesn’t appeal to business customers because it doesn’t have a keyboard which makes it not a very good e-mail machine….”

The market research industry needs to assert its expertise to differentiate between situations. In situations that involve volatility precision would demand a range of options rather than a precise estimate. It might be better to be approximately right rather than exactly wrong. 

The French and the British concurred, the reputation of the market research industry has been redeemed and I made a decent sum on my wager. My values, dictated by my small town lower middle class upbringing, makes me equate my financial gain to gambling which is a huge no no. I have hence decided to donate all my "earnings" to charity. 

Is the Market Research Industry – Unconsciously Conscious or Consciously Unconscious of Gender Bias

My hypothesis is that the Market Research Industry is not insulated from  gender bias but the question is whether it unconsciously conscious or consciously unconscious. Let me try and unpeel it from my perspective.

I have never understood why you cant generate random number through deterministic means but I have learnt how you can use deterministic means to combine random thoughts to create a new concept or idea. It’s the “corporate” version of the theory of dialectic materialism at work. So here are three random events which have coalesced into my thoughts summarised in the title of the blog.

Random Event 1 -  Sports is an area that is littered with examples of gender bias. For starters men and women are not paid the same – remember the row over the wages of men and women in tennis. Look at the photographs of the finish line of a marathon and you will find it littered with pictures of men even though women make up a significant chunk of the participants.

Last year before the cricket World T20, the English teams, men and women, went to meet the Prime Minister David Cameron (he was still PM then!!!). After the meeting, the teams went to the airport to fly out to India. The men flew business class while the women flew coach. It did make me realise that gender bias was still real, cuts across geographies, cuts across developed vs. developing economies and cuts across occupations and industries.  

Random Event 2 – Serena Williams recently announced that she was pregnant and within minutes the Twitterati figured out that she had won a Grand Slam when she was 8 weeks pregnant - women are clearly “super achievers”.

On my last holiday to India I met with an old friend. She is an alumnus of IIM Ahmedabad (the Harvard Business School of the East) and holds down a high pressure job. She does a full days work, comes home at 8PM and then cooks a hot meal for her family. I don’t recall a single conversation with any of my male batchmates about the need to come home and cook. Yet both genders have the same job specs and same expectations at the workplace. Stereotype or Prejudice?

Random Event 3 – I recently received an invite to attend a Market Research Conference. I looked through the event guide and noticed that a vast majority of the speakers were men. Surely the gender balance in the industry was better than the speaker panel. Why does it matter who speaks at a conference.

At a wedding I attended recently all the speeches were made by men. So I did a bit of research and found that the following is the order of speeches at a wedding

1.      The Master of ceremonies introduces and welcomes everyone

2.      Toast to the Bride and Groom made by the Father of the bride

3.      Response by the Groom and toast to the Bridal part

4.      Response on behalf of the Bridal party by the Best Man

5.      Toast to the Brides parents by the Best Man

6.      Response by the Bride’s Father

7.      Toast to the Grooms Parents by the Father of the Bride

There it is – speeches at the wedding is a “mans domain”.

The “post positivist” part of me says - surely you hear about bridesmaids now making speeches. Well again I did some research and while I did not find any study on the incidence of the bridesmaids  making speeches I found pointers to the fact that it is a small proportion. Here are the pointers

1. On Twitter, #bestmanspeech had 19 times as many mentions as compared to #bridesmaidsspeech. This would indicate that about 5% of weddings had a bridesmaids speech.

2.  Google trends, which is a good indicator of “needs”, shows an index of 40 for “best man speech” and 8 for “bridesmaids speech”. That would approximate to bridesmaids speeches at 20% of weddings.

A quick poll on LinkedIn coupled with speaking to a few people suggests, a number of 15-25% for the proportion of weddings with a speech from a bridesmaid, seems to range between “higher end of possible” to “optimistic” – clearly a minority and a small one at that.

“Traditions” tend to have a tenacity for survival and epitomise biases that we have inherited over generations. Speaking at weddings, arguably the most sacrosanct societal event and steeped in tradition, is a “mans domain”. It epitomises the importance of speaking at public functions (including conferences).

As someone who has been in market research for two decades the principle of epistemology, in particular the need to draw a distinction between justified belief and opinion, is well ingrained in me.  

First up, what is the gender balance of the industry. My friends at TNS pulled out statistics based on LinkedIn Pylon data showing male-female split in the market research industry is 55%-45%. That seems pretty good.

So was the conference I got invited to an aberration. I looked up the websites of major market research conferences and combed through the photographs of the speakers. Across 16 conferences, the average gender ratio was 63%-37% meaning women were under represented. To reassure you I looked at the median and it was 62%-38% i.e no different from the mean. To be fair, and in full disclosure, the last MRS in the UK had a perfect gender ratio i.e 50%-50%.

Basics of market research dictate that we need to “de-average” the data. Yes we do and I did. There is a (slight) difference between conferences in the D&E (67%-33%) and D (60%-40%) worlds but going back to the sports example economic development does not preclude discrimination (or stereotypes).

I looked up 1000 qualitative researcher profiles on LinkedIn (not a census but a reasonably large sample size) and found 19% were men i.e this segment of the industry is dominated by women. To be honest, from personal experience I would have expected the proportion of men to be in single digits but being a constructivist I accept the findings.

In a segment of the industry dominated by women, you would expect, the gender composition of speakers to be more reflective. Surely the feminist critiques of anthropological masculine bias would have impacted a woman dominated segment of the industry. I looked at qualitative specific conferences and it is really amazing that only 47% of all speakers were women i.e under representation of women is worse in the qualitative research segment of the industry.

	Male	Female	Female Representation Index
Average Market Research Conference	63	37	82
Qualitative Research Conference	53	47	58

You would expect that as we become conscious of our prejudices we would act to counter them. But maybe our prejudices always seem so rational. Is it a prejudice or a stereotype or is it discrimination? To quote Karl Marx “Necessity is blind until it becomes conscious. Freedom is the consciousness of necessity” 

American Opinion Polls are Rigged..............Nah But They are Wrong

With a week to go before the US Presidential elections, Real Clear Politics poll of polls has Clinton leading Trump by 2.2%. But scratch the surface and you see that the margin is depleting. In the last 5 polls published 4 are within the margin of error meaning technically they are a “dead heat”.

Here is the thing, I think the opinion polls are going to be proven wrong. Either Clinton will win by a big margin i.e well above 5% or Trump will win. Either way it wont be “too close to call”.

Here are five reasons why the polls will be proven wrong

1. Opinion polls have proven unreliable around the world – The polls got it all wrong in the elections in the UK, in Israel and just last week in Iceland. In Iceland, the anti establishment Pirate party was predicted to win but it did’nt. If you are looking for alternatives as a signpost here are two

(a) The UK bookies correctly predicted BREXIT. Currently at Ladbrokes, the split of bets is 80-20 in favour of a Trump Win.  

(b) In line with the opinion polls, the search for “move to Canada” which peaked when Trump won the primaries has surged in the past 7 days. Similarly, the search for the two candidates is now in a tie.

2.  American Polls are faulty to start with – The sample sizes of the polls hover around 1100. A poll of that size will surely struggle to represent the diversity of the US. Indeed the New York Times correctly pointed out how one 19 year old black man in Illinois was skewing the poll results. Essentially his demographic is  under represented in the panel and hence his voting intention (to vote for Trump) was upweighted significantly !!! Does it reflect the thinking of all of his demographic?

3. People lie to pollsters – That’s unfair to people who take the time to answer questions in a survey but it fair to say that when answering a human being “political correctness” plays a role. The polls conducted over telephone (Clinton winning) have a different response to those conducted online (Trump winning). Political correctness would explain the difference in the two polls. Remember on election day its just “you and your ballot in complete anonymity”. What role will political correctness play on the 8^th of November is something polls cant factor in.

4. Will they vote? – One problem that all pollsters wrestle with it is determining who will actually show up on election day. Will the Clinton supporters come out in droves now that “its not a done deal”. Will the independents who would otherwise have voted for Clinton stay at home? The outcome of the election could well nigh depend on the voter turnout and how this splits across demographics. 

5. Mood of the electorate” – You would expect voting to be a rational decision but since it is human beings casting their ballots emotions do play a huge role. It has been proven that the “mood of the electorate” can impact the outcome of the elections. Things like sporting events could impact this “mood”. With the rise of the social media we live in a “there are no lies, it’s a matter of perspective” world. So it is quite likely that the mood will be impacted by “news” or even “meme-able” headlines. Will there be more emails or Russian connection between now and next Tuesday?

Why Roy Hodgson will determine the outcome of BREXIT

Why Roy Hodgson will determine the outcome of Brexit. And why all opinion polls should be ignored.

Living in the UK, one cannot but follow the “BREXIT” debate with fascination. Both sides of the debate strongly put forth arguments around why the economy will be better off under their version of the future. So listening to the remain camp I am led to believe that the economy would thrive staying within the EU. Listening to the leave camp leads me to be believe that the economy would thrive by leaving the EU. Of course all arguments are backed by “statistics”. Reminds me of what Mark Twain once said “there are lies, there are damm lies and there are statistics”. 

It seems like it will impact everything – the economy, employment rates, immigration, the value of home prices, possibility of war and indeed the future of curry houses. The only thing that the outcome wont impact seems to be the chances of finding a vaccine for AIDS.  Thankfully, I don’t have the right to vote on the topic. I would find it impossible to arrive at a rational decision based on the “facts” being put forth by both sides.

I don’t find it surprising that this reflects in the split opinion among the electorate. While the latest opinion poll on Brexit shows “remain” to be ahead, I would interpret it as being a “dead heat”.  Indeed, looking back at the response to the question from 2013 shows remain and leave almost equally matched with a large proportion of population being “undecided”.

Source : https://en.wikipedia.org/wiki/Opinion_polling_for_the_United_Kingdom_European_Union_membership_referendum

Of Sportsmen and Gladiators

Sporting events are known to be associated with immense passion and ones involving national teams invariably raise the levels of nationalistic pride. Witness the euphoria surrounding the performance of Team GB at the London 2012 Olympics.

Andrew Healey from Loyola Marymount University showed that sports results can swing the outcome of elections. He showed that a win for the local college football team within 10 days of the election positively impacted the chances of the incumbent. The impact was stronger if the team had a strong fan base and were the underdogs.

Why is this relevant to Brexit? English voters account for almost 80% of the total voting population. The next largest block are the Scots (~10%) and they seem to be firmly in the remain camp (58% remain vs. 19% leave). So the English voters are likely to make the most difference.

With 3 days to go to the referendum, England play their last game of the league stage of Euro 2016. England have a large fan base and are clearly underdogs – a win would uplift the spirit of the nation and propel the remain campain forward.  Equally failing to reach the next stage, particularly on account of Wales (also in the same group) would be a dampener on the spirit of the nation and propel the leave campaign forward.

Is any opinion poll accounting for this? No. Well so how accurate are they going to be???

“The economy, stupid” – Campaign slogan of Bill Clinton

A recent ComRes telephone poll for The Daily Mail and ITV News shows that the economy is the largest factor influencing voters. It was for the first time that the economy was cited ahead of migration.

I can empathise with that given all the discussion around the potential impact on the prices of houses. Again depending on whom you listen to it could impact house prices negatively by upto 18% and mortgage rates would rise (Remain camp) or have no impact at all (Leave campaign). The more evocative topic is around jobs and business prospects. The debate was exemplified when the fate of curry houses was dragged into the debate!!!

Economists have studied approval ratings for the European project against the monthly unemployment rates. They found that when unemployment is rising approval of the membership goes down. Hardly surprising I would say. The correlation is more intuitive than the one between divorce rates in Maine and the per capital consumption of Margarine. (Source - http://tylervigen.com/spurious-correlations).

Here is the thing – the next tranche of employment data is due to be published on the 15^th of June. With 8 days to go before a major referendum we will have new information that will swing opinions of the electorate in general and the “14% undecided” in particular.

Is any opinion poll accounting for this? No. Well so how accurate are they going to be???

“In everything one thing is impossible : rationality” – Friedrich Nietzsche 

As human beings we are far less rational than we would like to think we are. Indeed as Robert A. Heinlein said “Man is not a rational animal, he is a rationalising animal”. Behaviour of human beings while voting is no different. Voting decisions are not based on a rational and objective reasoning despite their importance to the functioning of a democracy. Like with a lot of our other decisions, its all about emotions.

The outcome of Brexit will be determined by emotions. The feel good factor of either England, the quintessential underdogs, progressing to the next round or good employment figures would lead to a vote to remain. None of the opinion polls are factoring in emotional sentiments. Hence, they are at best “broad indicators”. The current “broad indication” is that it will be a close race – a conclusion I could reach by reading the newspapers and tuning into the radio.