Internet thinking vs Big data thinking: seeking common ground

What trends and strategies are more effective when we are marketing in the O2O era? There are two different ways of thinking, one is Internet thinking and the other is big data thinking.

Internet thinking and large data thinking have intersection but do not coincide. The current hype of the internet marketing case, basically stripped of large data, more is the theme of speculation and dissemination of the way hype. And big Data marketing is not limited to the Internet, it also includes offline marketing.

The debate on marketing art and science

How do you think of these two kinds of marketing thinking? In fact, the internet thinking and large data thinking PK, the essence is about marketing art and science controversy. One genre believes that marketing is an art, unutterable and inexpressible; another genre treats marketing as a scientific approach, and through the collection and analysis of consumer behavior data, the strategy of optimizing marketing is obtained.

Internet thinking can be understood as three keywords-experience, topic, communication. Experience is a consumer experience in the use of products or enjoy services, the Internet media can quickly transform the experience into the topic of dissemination, after the spread of the new experience, which leads to more topics and dissemination.

The big data is actually the science-oriented natural evolution of marketing. Large data thinking has three dimensions--quantitative thinking, related thinking, experimental thinking.

First, quantitative thinking, which provides more descriptive information, is the principle that everything is measurable. Not only sales data, price these objective criteria can form large data, even customer sentiment (such as color, spatial perception, etc.) can be measured, large data contains all aspects related to consumer behavior; second, related thinking, everything can be linked, consumer behavior of different data are intrinsically linked. This can be used to predict consumer behavior preferences; third, experimental thinking, everything can be tried, the information brought by large data can help to develop marketing strategies.

This is the three large data application level: First is the description, then the prediction, and finally the introduction.

Everything is measurable: Disney MagicBand Bracelet

The company recently invested 1 billion of dollars in offline customer tracking and data collection to develop a magicband bracelet. When visitors wear the hand ring with the function of location collection, the park can understand the distribution of the tourists in different areas through the positioning system, and tell the visitors the information, so as to facilitate the tourists to choose the best play route. In addition, the user can also use the mobile booking function, through the positioning of the hand ring, feeding staff can deliver fast food to the user hands. The use of large data not only improves the user experience, but also helps to ease the flow of people in the park. The collected customer data can be used for precision marketing. This is an example of all that can be measured, and offline activity can also be measured.

Everything can be connected: online Booking tracking System

An internet company that makes a meal delivery, installs a software and tracking system on the bicycles and cars to be sold, collects a lot of data from the distribution, such as who ordered what to sell, after what route, to whose home ... and through the analysis of the data, you can draw what restaurant is more popular, the quickest path is that one, on this basis for the merchant to provide preparation recommendations, and planning a reasonable and efficient delivery route. By analyzing the seemingly unrelated large data on the surface, companies can provide value-added services that optimize restaurant operations.

All can try: Electric Business page recommend function

The other product recommendation of the commodity page is an important function (for example, "the person who bought the product has bought XXX"). How to quantify and optimize the effectiveness of the recommended function? A research institute has done such a test: in order to recommend to the user all/shielding part of the recommendation/shielding all recommendations, after one months of testing, tracking the purchase of the subjects, found that not shielding the recommended short-term effect of the highest purchase volume. and shielding all the recommended effect is better than the shielding part of the recommendation. Consumers who had previously purchased the goods were shielded from the recommendation, and the sales of the goods fell faster, thus leading to a greater role for loyal customers. More interesting is the long-term effect of the recommended feature. The study found that, regardless of whether the user purchased the recommended product during the first purchase, the second visit to follow this rule: not blocked recommended customers, 10% of the people will visit again, the recommended access rate is 9%, and the actual conversion to the number of visits is 8%, if combined with the old customer recommendations will be better, In the end, more than 10% of the revenue increase is generated. Overall, the recommendations are more effective.

From the description to the forecast, then to the production strategy

Social network analysis and tracking, the frequency of keywords on the consumer social network into visual expression, classification of consumers, and then to target customer base for precision marketing, which is the description of large data marketing stage.

The prediction phase is a case study of credit card usage. Originally each bank can only see the consumer's bank card record, the bank according to the consumer record to reward customers. The problem is that the consumer's use of non-bank credit cards is unknown, banks are unable to understand the actual consumption of their customers and which are hidden "big spenders". The difficulty in tackling the problem is that his data records are difficult to obtain, so the research institute uses data from third party retailers to build models, The two kinds of data are fused and the actual consumption of consumers is forecasted. Consumers, who were likely to spend only 2000-3000 yuan a year in the model, actually consumed 40,000, and these people became very potential bank customers.

In the introduction stage, the Bank can adjust the customer's reward policy according to the forecast result, for example, to increase the return point to the customer who consumes 3000 yuan per year, or to provide richer product point exchange products, so that this part of the crowd becomes the loyal customer of the bank.

Oriental wisdom and Western knowledge neglected

How does Internet thinking pk big data thinking? The word "Internet age" is particularly hot in China, but it has not been heard in the United States. This is because the Internet thinking is more consistent with the traditional oriental way of thinking. Oriental culture emphasizes wisdom, while the West emphasizes knowledge, wisdom comes from experience, and knowledge comes from data. Zhuge Liang and Yi are representative of a typical group of wisdom PK knowledge. Yi is the biggest opponent of Zhuge Liang, who may be the best user of the early big data. From Zhuge Liang to sleep, eat a few bowls of rice, he can judge Zhuge Liang live not long, and Zhuge Liang by virtue of wisdom guess Amat timid, dare not enter empty city. Chinese people advocating wisdom, may pay more attention to internet thinking, but the internet thinking is not enough, but also to the data have a deeper understanding and better use.

Big Data thinking is not as exciting as Internet thinking. A recent study shows that companies with large data are averaging 6% more than companies that do not use big data. 6% may not be so humble, but "add up, Mickle," in the fierce competition environment, this is to enable enterprises to survive, stand out from the capital. Of the top ten U.S. electric dealers ' websites, 8 are traditional retailers, and only 2 are All-electric (Amazon and Shellfish). Traditional retailers have a lot of data--Wal-Mart's data is at a petabyte a day, and the data resources can be transformed into an enterprise's stamina to win games. Due to the large data age there is intrinsic to the feedback logic from the enterprise to the strong, enterprises will produce more data, the understanding of consumers will be more profound, more accurate marketing, enterprises become stronger, and then produce more data to form positive feedback, which is a final data-driven growth model.

Using large data to guide marketing decision-making is the intrinsic logic of many mergers and acquisitions strategies.

The ideal state is the combination of science and art. Wearable motion camera manufacturer GoPro's listing is a successful case of combination of big data thinking and internet thinking. The company, which originally produced only physical cameras, first developed a camera with WiFi, users can instantly share photos and video to the Internet, the inherent logic is from experience to spread to share the internet thinking; after that, GoPro into the analysis and application of large data to classify the contents of the user's shooting. To match the content with potential advertisers. In addition, GoPro also bought television channel broadcast rights, through the data analysis which time is suitable to play what content, then matches with the advertisement, achieves the accurate marketing. GoPro has developed a social platform, even a media function, from a physical camera manufacturer.

Chinese people have marketing wisdom, enterprises through the use of large data and the combination of the internet era of the giant, is very likely to achieve.