Information Technologies RSS

  Using Big Search Data to Map Your Market  Premium

Skiera, Bernd; Ringel, Daniel M.
Markets today are flooded with an increasing number of products and brands, making it difficult for companies to track how their products compete in the market. Big search data allow companies to obtain valuable insights on market share as well as on key competitors. But collecting and analyzing vast amounts of data pose huge challenges. In this article, the authors describe how they used clickstream data to visualize competition in product categories containing more than 1,000 products. Both manufacturers and retailers can profit from their approach, which yields more meaningful information on customer behavior than traditional market research methods.


Recently, we ran a marketing workshop with an experienced business development manager from Amadeus, the tech provider for the travel industry. The workshop focused on Travel Audience, Amadeus' online advertising platform for travel companies. Participants were given four data sets, ranging in size from 770,000 to more than 9 million "observations" -- which reveal how users or consumers behave and interact in a specific market category. We asked the participants to analyze the data and answer a few basic marketing questions: Where does the company make the most profit? Where should it spend more money on advertising? What is the return on investment in terms of advertising?

To our surprise, the participants were stumped. First, it took them a very long time to convert the data into a format that would allow them to answer basic marketing questions, because the sample sizes were far too big for a conventional spreadsheet program. They struggled with separating causality from correlation. And finally, they complained about the amount of data they had to analyze -- until we reminded them that the data sets they had been given barely covered a week!

This experience illustrates that, unlike yesterday's marketing professionals, marketers today need to upgrade their data-handling skills. It's not enough to ask a junior professional to do the work for you. Marketers who can pull and read the numbers themselves are in a much better position than those who have to run to someone else each and every time they need to retrieve key data, waiting days or weeks to get the information they want from the IT department. Being able to read and react faster to market shifts will lend a competitive advantage.

With so many brands and products and so much big data available today, where does one begin? Interpreting the competitive relationships among 1,000 different products is extremely difficult and would entail a gigantic, unwieldy table. As such, we came up with a way of visualizing competitive market structure in a single map. This makes it more manageable to devise effective competitive strategies, pricing policies and communication.

This article explains our approach. Using data of large markets with more than 1,000 competing products, we show how companies can tap into big data and visualize competitive market structure. Various parties can benefit from this method, which reveals valuable insights on customer behavior.

The Big Learning Curve of Big Data
Ever since the advent of the browser, marketers have been trying to get to grips with the information boom. The internet enabled marketers to measure, on a grand scale, when an advertisement was shown and how consumers reacted to it -- how many clicked and how many proceeded to purchase.

With the ubiquity of increasingly affordable and powerful smartphones, the sea of data has grown exponentially. On top of all existing data, companies can now also collect vast quantities of location-based information on consumer behavior around the clock.

While having more and better information is a good thing, big data present distinct challenges -- from poor data quality, to incomplete data sets, to mistaking correlation for causality.

There are several signals that marketers are behind the learning curve. It used to be that the chief concerns of marketing were to develop creative slogans and strategies. In fact, many people went into marketing precisely because they didn't want to do math. The digital revolution, however, has changed all that. While coming up with creative content will always be central to the profession, marketers now have to take a data-driven approach to prove that their ideas actually work.

This development doesn't mean marketing professionals have to become IT specialists or set up databases and IT architecture themselves. Such tasks can and should stay with the IT department. However, it does mean that successful marketers will be those who also have solid math and data-management skills. The rise of crowdsourced platforms, such as Kaggle, to carry out predictive modeling and analytics, as well as the growing popularity of online data-analysis courses, are testament to marketing's wider remit.

Market Share and Competition
Recently, a manager at a price-comparison website mentioned that his firm could quickly predict the success or failure of a new product by simply following the outbound clicks to retailers' websites. That gave us an idea: why not use clicks to predict market share?

We also decided to do something that market research firms don't do very well. Traditional approaches are largely based on scanner panel data that require multiple purchases within a product category to detect switching between products as an indicator of their substitutability. We sought to identify competition among durable products that consumers typically buy just once. Market share and competition are two important areas in which big data can provide key insights for marketers.

To continue reading, you need to SUBSCRIBE or PURCHASE the article through IESE Publishing. Remember, if you are a Member of the IESE Alumni Association, you can access all the contents of the magazine as part of your membership, using the same username and password that you normally use to log in as a Member.
This article is based on:  Using Big Search Data to Map Your Market
Publisher:  IESE
Year:  2017
Language:  English
Print
Buy it English
Buy it Spanish
Share