Why are we still blind in the big data era?
The author of the material explains why public data is not always effective for improving business processes. She bases her arguments on examples of American companies from the retail and e-commerce industries.
We live in a world where we are scattered every day with scientific data on everything from stock prices, market capitalization, retail sales and consumer spending, to analyzing price jumps in cryptocurrencies and the impact of innovation on the future of payments, commerce and financial services. But even despite all this abundance, we simply act blindly when it comes to the really important things for business and consumers.
We act blindly, because we do not have the data necessary to construct and put into practice schemes and models that would allow us to make verified and balanced decisions that can direct our businesses and even the economy as a whole to the right direction.
We were lost in that saturated data stream which surrounds us.
And the evidence of this state of affairs we observe every day.
Reporters are alarming: a strong economy and historically high unemployment levels cause people to pull out their wallets and buy more than ever.
Indeed, things in physical retail go very well, despite all the talk about his dark fate .
And yet, according to Fung Global Retail & Technology, 6985 stores closed in 2017, which is 229% more than last year, and certainly much more than in 2008, when this trend first became noticeable. The number of bankrupts in retail increased by 30% , and the list was supplemented with new familiar names, such as Payless ShoeSource, Toys R Us and The Limited.
At the same time, the market capitalization of the 20 largest retailers has fallen by more than $ 230 billion over the past two years. Mall operators also show poor results, since the key stores in these stores are overloaded with goods and not generating income for their brands, casting doubt on the very existence of such shopping centers. Analysts even predict that every fourth out of 1,100 moles operating today will close by 2022, that is, after four years.
At the same time, judging by data from the US Census Bureau, there is nothing wrong with physical retail, since more than 90% of all retail sales take place here. Many analysts use this information as a counter-argument for claims that things go wrong in physical retail .
All this talk about e-commerce is too bloated, they say, pointing to data that, at first glance, seems to suggest that everything is good in the country of physical retail.
And they immediately add that the slow growth of “physics” is explained by the presence of an already huge customer base in this area, while the growth of online is due to the fact that it has not yet gained really large scales.
However, these arguments can hardly console 7 thousand stores closed last year, 700 of which declared themselves bankrupt. They will not help the network of department stores, which closed a total of 550 outlets last year.
And besides them, books, music stores, stationery companies, sporting goods and clothing stores, whose store sales go down as more and more consumers are switching to online purchases, feel the seriousness of the situation.
The problem with the Census Bureau data is that they record averages, but this approach cannot be called effective in practice. The average February temperature in Boston fluctuates at 3 degrees, but on the last Saturday it was 11 degrees, and a week ago - minus two.
Two years ago, bureau officials admitted (for the first time ever) that their accounting systems for collecting and analyzing data were not set up to track transactions in the digital world, increasingly helping to erase the lines between online and offline trade and allow manufacturers to sell their products directly to own brand And because the Bureau simply did not take into account these transactions.
The situation is aggravated by the fact that many retailers who submit their reports to the Bureau are unlikely to “bother” with the exact transfer of information about such processes as buying online with the subsequent self-removal of goods in the store.
That is, it turns out that the Bureau’s data contain understated figures for the share of online sales in retail. In addition, there remains an open question about the accuracy of this data.
So it turns out that quarter after quarter we live with these contradictions, pretending that everything is in order, while traditional retail is slowly going to the bottom.
Another interesting observation is that even if the current Census Bureau data are correct, the forecast for online sales growth and a slowing average sales growth for physical retail will result in parity between them in about 30 years.
However, for sure it will happen faster and earlier than the Bureau reports.
Another topic concerns perennial violent disputes among economists about whether GDP can be considered the most effective way to change the welfare of an economy.
Most of the GDP is provided by consumer spending of the population, determined by multiplying the price by the quantity of goods purchased. This approach was invented when the world economy was more focused on industrial production, and the products and goods produced, respectively, provided most of the sales.
Consumers at that time always paid a certain price for receiving goods and services.
However, the Internet, mobile applications, smartphones and progress in network bandwidth have changed the understanding of the essence of production, and with it some business models that make possible a commodity-money turnover between suppliers and consumers.
Today, more than two-thirds of US residents use a smartphone, which is twice the figure seven years ago. According to eMarketer, by the end of 2018, no less than a third of the world's population, that is, more than 2.5 billion people, will have smartphones.
As for industrial production, today it accounts for about 11.7% of total sales and sales, which is 25.4% less than in 1947.
The combination of mobile applications, phones and technologies such as GPS inspired entrepreneurs of all stripes to create new or enhance existing digital intermediaries that unite two groups of participants interested in them. In some such cases, we are talking about new businesses based on virtual platforms, but the underlying business models have been tested by time for 3 thousand years.
All this makes it obvious that standard methods of measuring GDP completely ignore the economic benefits that consumers receive from free offers.
As an example, Venmo (as a special case of P2P) is a free service for both the sender and the recipient. This feature leads to the fact that the economic value of simplifying the transfer of money between the two sides simply passes by any metrics: for economists, this is just one big fat zero. As for the attempts to monetize these services, especially considering their effectiveness, this topic is especially painful for banks and PayPal.
Similar issues are even sharper when it comes to digital content platforms, from Facebook or Google to television networks that are monetized by advertising.
David Evans, an economist and chairman of the Global Economics Group, wrote in his recently published research paper that in 2016, adult Americans spent 437 hours consuming content on existing media advertising. Of course, consumers thought such a pastime is quite valuable, otherwise why would they continue to be on these sites? And even if the benefits obtained during the visit to these resources could be assessed with the help of the minimum salary, then its cost would be at least $ 2.8 trillion.
At the same time, the lack of a qualitative way of interpreting the data that we have (whatever they may be) can lead to market imbalance and other far-reaching consequences.
Two weeks ago, eBay decided to take the payment process on the platform under its own control, which led to a sharp drop in PayPal shares within 48 hours. A careful analysis of the facts, however, showed that now PayPal has nothing to worry about, since the transition to new payment solutions will require time and effort from eBay.
Such situations occur regularly whenever Amazon announces the expansion of its activities and beyond the traditional framework of retail or e-commerce, even despite the fact that, according to the Census Bureau, sales in e-commerce are very small. And we know from other sources that Amazon accounts for half of this “small” market.
Recently, for example, the company announced the launch of its own delivery service, as a result of UPS shares with FedEx lost $ 25 billion of their total market value (UPS lost $ 18 billion, and FedEx lost $ 7 billion).
A few months earlier, Amazon announced it had applied for a license to wholesale pharmaceutical products in 12 states, making it possible for other players to understand their plans to enter the pharmaceutical market, which is estimated at $ 560 billion. Shares of each of the largest market operators Express Scripts, Caremark and Optimum then also fell by 4% , even despite the fact that Amazon is new to the market unknown to the company, the true extent of which is known only to its players.
And when Amazon acquired Whole Foods in August last year, Kroger's market capitalization fell from $ 30 billion to $ 18 billion in September 2017. And although Kroger shares have since increased, market capitalization still has not returned to its previous level, $ 24 billion. This is despite the fact that the grocery business in the US is very segmented and the total share of Amazon and Whole Foods in the market, according to 2016, is less than 3 percent.
But here we again run into the question of the applied effectiveness of the data collected and their analysis, which, as we have already learned, varies from case to case. What really has a significant impact on the market is the tangible impetus that Amazon has led to new retail segments, regardless of the fact that, according to the Census Bureau, almost all retail sales take place in offline stores.
Topics such as big data, artificial intelligence and machine learning are very popular today. Often, cases of processing with their help huge amounts of data flooding the information space are discussed.
It is estimated that humanity generates 2.5 quintillion bytes of data daily. Quintillon is a million trillion. In other words, quite a lot.
However, instead of “feeding” all this wealth to machines, we may need to spend more time creating well-thought-out models and approaches that can help formulate the most important issues related to innovation and only then go looking for the data necessary to answer them.
Indeed, in the absence of such an approach, the data simply turn into sets of numbers.
We live in a world where we are scattered every day with scientific data on everything from stock prices, market capitalization, retail sales and consumer spending, to analyzing price jumps in cryptocurrencies and the impact of innovation on the future of payments, commerce and financial services. But even despite all this abundance, we simply act blindly when it comes to the really important things for business and consumers.
We act blindly, because we do not have the data necessary to construct and put into practice schemes and models that would allow us to make verified and balanced decisions that can direct our businesses and even the economy as a whole to the right direction.
We were lost in that saturated data stream which surrounds us.
And the evidence of this state of affairs we observe every day.
Happy retail mask
Reporters are alarming: a strong economy and historically high unemployment levels cause people to pull out their wallets and buy more than ever.
Indeed, things in physical retail go very well, despite all the talk about his dark fate .
And yet, according to Fung Global Retail & Technology, 6985 stores closed in 2017, which is 229% more than last year, and certainly much more than in 2008, when this trend first became noticeable. The number of bankrupts in retail increased by 30% , and the list was supplemented with new familiar names, such as Payless ShoeSource, Toys R Us and The Limited.
At the same time, the market capitalization of the 20 largest retailers has fallen by more than $ 230 billion over the past two years. Mall operators also show poor results, since the key stores in these stores are overloaded with goods and not generating income for their brands, casting doubt on the very existence of such shopping centers. Analysts even predict that every fourth out of 1,100 moles operating today will close by 2022, that is, after four years.
At the same time, judging by data from the US Census Bureau, there is nothing wrong with physical retail, since more than 90% of all retail sales take place here. Many analysts use this information as a counter-argument for claims that things go wrong in physical retail .
All this talk about e-commerce is too bloated, they say, pointing to data that, at first glance, seems to suggest that everything is good in the country of physical retail.
And they immediately add that the slow growth of “physics” is explained by the presence of an already huge customer base in this area, while the growth of online is due to the fact that it has not yet gained really large scales.
However, these arguments can hardly console 7 thousand stores closed last year, 700 of which declared themselves bankrupt. They will not help the network of department stores, which closed a total of 550 outlets last year.
And besides them, books, music stores, stationery companies, sporting goods and clothing stores, whose store sales go down as more and more consumers are switching to online purchases, feel the seriousness of the situation.
The problem with the Census Bureau data is that they record averages, but this approach cannot be called effective in practice. The average February temperature in Boston fluctuates at 3 degrees, but on the last Saturday it was 11 degrees, and a week ago - minus two.
Two years ago, bureau officials admitted (for the first time ever) that their accounting systems for collecting and analyzing data were not set up to track transactions in the digital world, increasingly helping to erase the lines between online and offline trade and allow manufacturers to sell their products directly to own brand And because the Bureau simply did not take into account these transactions.
The situation is aggravated by the fact that many retailers who submit their reports to the Bureau are unlikely to “bother” with the exact transfer of information about such processes as buying online with the subsequent self-removal of goods in the store.
That is, it turns out that the Bureau’s data contain understated figures for the share of online sales in retail. In addition, there remains an open question about the accuracy of this data.
So it turns out that quarter after quarter we live with these contradictions, pretending that everything is in order, while traditional retail is slowly going to the bottom.
Another interesting observation is that even if the current Census Bureau data are correct, the forecast for online sales growth and a slowing average sales growth for physical retail will result in parity between them in about 30 years.
However, for sure it will happen faster and earlier than the Bureau reports.
What are the innovations?
Another topic concerns perennial violent disputes among economists about whether GDP can be considered the most effective way to change the welfare of an economy.
Most of the GDP is provided by consumer spending of the population, determined by multiplying the price by the quantity of goods purchased. This approach was invented when the world economy was more focused on industrial production, and the products and goods produced, respectively, provided most of the sales.
Consumers at that time always paid a certain price for receiving goods and services.
However, the Internet, mobile applications, smartphones and progress in network bandwidth have changed the understanding of the essence of production, and with it some business models that make possible a commodity-money turnover between suppliers and consumers.
Today, more than two-thirds of US residents use a smartphone, which is twice the figure seven years ago. According to eMarketer, by the end of 2018, no less than a third of the world's population, that is, more than 2.5 billion people, will have smartphones.
As for industrial production, today it accounts for about 11.7% of total sales and sales, which is 25.4% less than in 1947.
The combination of mobile applications, phones and technologies such as GPS inspired entrepreneurs of all stripes to create new or enhance existing digital intermediaries that unite two groups of participants interested in them. In some such cases, we are talking about new businesses based on virtual platforms, but the underlying business models have been tested by time for 3 thousand years.
All this makes it obvious that standard methods of measuring GDP completely ignore the economic benefits that consumers receive from free offers.
As an example, Venmo (as a special case of P2P) is a free service for both the sender and the recipient. This feature leads to the fact that the economic value of simplifying the transfer of money between the two sides simply passes by any metrics: for economists, this is just one big fat zero. As for the attempts to monetize these services, especially considering their effectiveness, this topic is especially painful for banks and PayPal.
Similar issues are even sharper when it comes to digital content platforms, from Facebook or Google to television networks that are monetized by advertising.
David Evans, an economist and chairman of the Global Economics Group, wrote in his recently published research paper that in 2016, adult Americans spent 437 hours consuming content on existing media advertising. Of course, consumers thought such a pastime is quite valuable, otherwise why would they continue to be on these sites? And even if the benefits obtained during the visit to these resources could be assessed with the help of the minimum salary, then its cost would be at least $ 2.8 trillion.
Chatter about data
At the same time, the lack of a qualitative way of interpreting the data that we have (whatever they may be) can lead to market imbalance and other far-reaching consequences.
Two weeks ago, eBay decided to take the payment process on the platform under its own control, which led to a sharp drop in PayPal shares within 48 hours. A careful analysis of the facts, however, showed that now PayPal has nothing to worry about, since the transition to new payment solutions will require time and effort from eBay.
Such situations occur regularly whenever Amazon announces the expansion of its activities and beyond the traditional framework of retail or e-commerce, even despite the fact that, according to the Census Bureau, sales in e-commerce are very small. And we know from other sources that Amazon accounts for half of this “small” market.
Recently, for example, the company announced the launch of its own delivery service, as a result of UPS shares with FedEx lost $ 25 billion of their total market value (UPS lost $ 18 billion, and FedEx lost $ 7 billion).
A few months earlier, Amazon announced it had applied for a license to wholesale pharmaceutical products in 12 states, making it possible for other players to understand their plans to enter the pharmaceutical market, which is estimated at $ 560 billion. Shares of each of the largest market operators Express Scripts, Caremark and Optimum then also fell by 4% , even despite the fact that Amazon is new to the market unknown to the company, the true extent of which is known only to its players.
And when Amazon acquired Whole Foods in August last year, Kroger's market capitalization fell from $ 30 billion to $ 18 billion in September 2017. And although Kroger shares have since increased, market capitalization still has not returned to its previous level, $ 24 billion. This is despite the fact that the grocery business in the US is very segmented and the total share of Amazon and Whole Foods in the market, according to 2016, is less than 3 percent.
But here we again run into the question of the applied effectiveness of the data collected and their analysis, which, as we have already learned, varies from case to case. What really has a significant impact on the market is the tangible impetus that Amazon has led to new retail segments, regardless of the fact that, according to the Census Bureau, almost all retail sales take place in offline stores.
Topics such as big data, artificial intelligence and machine learning are very popular today. Often, cases of processing with their help huge amounts of data flooding the information space are discussed.
It is estimated that humanity generates 2.5 quintillion bytes of data daily. Quintillon is a million trillion. In other words, quite a lot.
However, instead of “feeding” all this wealth to machines, we may need to spend more time creating well-thought-out models and approaches that can help formulate the most important issues related to innovation and only then go looking for the data necessary to answer them.
Indeed, in the absence of such an approach, the data simply turn into sets of numbers.
Source: https://habr.com/ru/post/410687/