What does big data marketing do? How do you do it? How are you doing?

1. Who is doing big data marketing?

Everyone says they are doing big data marketing.

Based：

Business: and marketing-related, so the government Office unit data is not considered. Volume: has enough valuable data. This is a lot of internet companies and traditional large enterprises can meet. Technology: Technical ability to handle large data.

Above 3 factors, the domestic can enlarge the data marketing really only bat three. Foreign unfamiliar, here do not talk.

As we all know, the data characteristics of three are different. Tencent advantage in social data, Alibaba advantage in commodity and transaction data, Baidu advantage in the whole network information, consumer behavior and active demand data.

Of course, such as peace, the big companies such as Procter and Gamble, Wal-Mart, themselves must accumulate a large amount of data, based on the data mining, the past has been done in the Web analytics and other business, although now dubbed "Big Data", but this is not the same as the big data we discussed, They use traditional data tools to analyze the extraction of certain data, can be based on those data mining, but the number has increased, generally still belong to the traditional small data category.

2. What does big data marketing do?

Marketing Business Type

If a rough division, advertising marketing budget can generally be divided into effective marketing and brand marketing two large chunks, according to their own development needs and industry characteristics of the business have a focus. For example, in the past, such as the Beijing-east, a shop, such as electric business enterprises, such as peace and other financial enterprises mainly to do the effect of marketing, the Internet is more suitable for the effect of the media channel; such as FMCG customers, Mercedes-Benz Audi and other automotive customers mainly to do brand marketing, Now, of course, things are changing, mainly reflected in:

more and more brand advertisers are also starting to put the brand marketing budget on the internet to do more and more marketing forms more and more comprehensive. Effect class customers gradually began to do brand (Jingdong), brand customers also began to do the effect (car, check line 4S shop sales conversion).

2.1. Effective Marketing

Effective marketing, the Internet people too clear. As a result of business characteristics, the past Baidu and Alibaba large data mainly used by small and medium-sized customers and consumers personalized advertising, Tencent is mainly consumer-oriented personalized advertising (Ali can also use the payment data for credit risk assessment, but the financial aspects).

For example, as you are relatively familiar with large data training to optimize the data mining model, Amazon and other retail dealers generally apply this personalized recommendation technology, in my opinion is only one of the application types in marketing. All kinds of personalized search and display ads, including bat and major electric dealers, are the way. Basically is the actual effect marketing, examination CPC. In the eyes of many internet people, due to familiarity with the effectiveness of marketing, there will be a perception that marketing is these things. Hanging advertisement, examination Cpm/cpc/cpd/cps.

2.2 Brand Marketing

According to my observation, many internet people are actually unfamiliar with brand marketing.

Here need to say why do brand marketing first?

Reason 1. The need to make money: effectiveness (effect) marketing money to make the ceiling, the Internet media to rob Brand marketing big head of the budget.

The entire advertising market, large advertisers hand in the budget, the big head or brand marketing budget, the media on the traditional media (such as television, etc.), the majority of the money cast to the digital media only 10%~30% (approximate number).

Reason 2. Customer needs: Don't talk to me about CPC anymore, a lot of things can't be measured by clicks, how is the brand's popularity, reputation, loyalty measured by the CPC? It is therefore necessary to provide integrated marketing solutions to major customers.

Baidu used to be the typical representative of effect marketing.

There is a cognitive, Baidu in the eyes of netizens is a search tool, making money by SEM, by bidding rankings, earn bad advertisers money. I think this is also the mental laziness that @Fenng mentioned.

Baidu in addition to a wide range of SME customers, there are a large number of large brand customers, such as Procter and Gamble, Mercedes-Benz, BMW, Peace, l ' oreal and so on, a large proportion of income. For these big customers, the need for brand marketing.

So the value of the big Baidu data is reflected. The advantage of Baidu data is that the whole network information and consumer's real behavior and demand expression. The breadth of coverage is not comparable to commodity trading data. So it lays the foundation for the brand to provide integrated marketing solutions.

Based on the above, the most valuable way for Baidu is based on large data to provide brand marketing solutions.

3. What about Big Data marketing?

Basic process of marketing based on data

The basic marketing process of large data is not different from the basic process of data analysis in the past, and it needs to collect and process data, model and analyze data, and interpret data in three dimensions after defining business problems. But the impact of big data on three levels makes concrete practices different from tradition.

3.1 Data layer: collecting and processing data

Traditionally, the process of data acquisition is limited, conscious and structured for data collection, such as questionnaire research. The data you can collect must be what you can imagine. The structure of the data is better. General database MySQL and even Excel can meet the data processing process.

In the Internet age, the acquisition process of large data is essentially infinite, unconscious, unstructured data collection. All kinds of complex behavioral data are uploaded to the server in the form of behavioral logs. Tools such as Hadoop, MapReduce, etc. are not to be summarized.

3.2 Business Layer: Modeling analysis data

Using the data analysis model, for example, basic statistics, machine learning, such as data mining classification, Clustering, association, prediction algorithms, traditional data and large data practices are not very different, such as banks, communications operators, retailers have long been mature use of consumer attributes and behavioral data to identify risks and payment possibilities. However, due to the large expansion of data volume, the algorithm also obtains the space of great optimization and promotion.

3.3 Application layer: Interpreting Data

The most important part of data-directing marketing is interpretation.

Traditionally, after the marketing problem is defined, the corresponding data is collected, and then the data are analyzed according to the established modeling or analysis framework, and the hypotheses are validated and interpreted. The space for interpretation is limited.

and large data provides a possibility, not only according to marketing problems, closed to mining the corresponding data to verify, but also open to explore, come up with some may and common sense or experience judgment completely different conclusions. The point of interpretation becomes very rich.

4. How is the big data marketing doing?

Big Data Explore brand marketing How are you doing?

Alibaba how do I do not know, feel Ali Big data also mainly in the effect of marketing aspects. Through the "West Lake Achievement Big Data Summit," the report appears.

My department is based on the customer's real business needs and problems, through the application of large data to provide solutions, and customers have produced many excellent model methods and case results.

Similar to the traditional brand marketing, large data in the main application direction of brand marketing are:

consumer insight into media value Research market competition analysis brand management

Thanks to Baidu good contribution. The article comes from knowing, view original post please enter >>>