Small data strategy run-off big data age

I've heard from American colleagues about one thing: a client of an advertising company happens to be a Marine colonel who is talking about data reliability. The colonel said: "If I get an intelligence on the battlefield, even if I am not 100% sure of the accuracy of the information, I will still make decisions based on intelligence." "He strongly believes that intelligence is better than no intelligence-we cannot ignore the information simply because it is not reliable." Some would consider the colonel a "small data" supporter.

There has been a lot of discussion around "big data" recently, and large data is characterized by large numbers, speed and variety, and new data processing methods are needed. Enterprises want to use large data to optimize decision-making and improve operational efficiency, through data mining to better understand customer behavior and preferences, and even use large data to predict the stock market dynamics. Some companies have been successful with big data strategies, and other companies are starting to invest in the infrastructure, software, and talent needed to make big data.

However, there is a point to be explained. Many enterprises, probably most of them, are in a relatively low data environment, unable to obtain the vast amount of information needed for advanced analysis and data mining. For example, transaction data for sales terminals (POS) have not been standardized in emerging markets. In most business-to-business businesses, companies have access to their own sales and distribution data, but rarely understand overall market sales or the products sold by competitors. There are a limited number of potential customers in highly specialized or concentrated markets, such as auto parts suppliers. These enterprises have to be content with small data, even in the case of insufficient data or uneven quality of data, but also through the use of limited data to form insights.

Some commentators say big data is not just a new source of data and analytical technology. Big data radically changes the way that decisions are made, from management to making decisions based on intuition, to making decisions based on data. For those companies that lack complete or definitive market data, they must strive to efficiently use existing data (which may have insufficient data) or use innovative and Low-cost methods to obtain new data.

Let's look at an example of a small data strategy. A large beverage producer wants to boost its product's immediate drinking channel, namely, bars, restaurants and entertainment outlets. For years, the company has been buying synchronized data from the same institution, which covers 100,000 points of sale. However, these data are collected and combed to meet the general needs of a large number of customers, and the use of standardized partitioning methods does not help the beverage business understand how to effectively serve different market segments. As a result, the company decided to adopt a series of small data technologies to develop solutions based on demand.

The company first used observational research, visited bars and restaurants, and made a qualitative record of consumers and their consumption patterns. The company used the data to obtain a more workable definition of market segments. The next step is to quantify the market segments to identify how many sales points are in each segment. The company developed a formula based on observed characteristics and then asked salespeople to classify the bars and restaurants covered by the sales area according to the formula. (This is a typical small data strategy: the internal filling of data gaps.) Eventually, the company designed specific product portfolios, pricing, and marketing projects based on major market segments. The company has carried out pilot projects in two major cities, with overall sales and market penetration significantly enhanced, and is spreading the initiative nationwide.

Haier in China, for example, uses information gathered by service engineers to drive innovation. The most famous washing-potato washer was in the late 90, when engineers discovered that some rural consumers were clogging up with washing machines to wash vegetables, and Haier used the information to develop a new washing machine that cleans potatoes, sweet potatoes and peanuts in addition to washing clothes.

After clear thinking, all information can be used to enhance the product, customer experience or company profit. As a result, small data strategies can include any approach that allows companies to gain more customer insights at low cost. As shown above, mining small data does not mean investing heavily in data acquisition, hardware, software, or technical facilities.

In addition, businesses need to do three things:

Committed to developing evidence-based decision-making mechanisms. This is often the case when companies find themselves increasingly competitive or unable to accurately capture changing consumer habits and preferences. For market-oriented enterprises, the decision mechanism based on facts is an important source of gaining competitive advantage.

Have the will to do and learn. Since small data strategies do not need to go through third parties, companies have to make their own attempts and learn from their mistakes. Once a few priorities have been identified, a series of pilot projects will bring valuable experience to the enterprise, and those that have been successful through small data strategies can just inspire other companies.

Promote creativity. To get richer data, businesses need to promote creativity and naturally integrate innovation into their interactions with consumers. For example, retailers can use the ipad to complete a survey with their customers. Enterprises can also be in any information registration function of the site to ask consumer preferences questions to collect relevant information, further improve the site collected some of the basic data. Call center customer service and Consumer dialogue is also a good opportunity to gather information, conducive to more in-depth consumer insights. Some companies will also organize a group of mature consumers who are good at digging and listen to the suggestions of the user group in the process of developing new products. Some companies rely on sales reps for information about consumer preferences and competitor activity. But more importantly, companies must devote more effort to collecting and interpreting the data that has been generated.

Businesses typically pick a product, a region, and a problem that requires attention, and start a data analysis tour, and one or more pilot projects are carried out. Executives will prove to themselves and other members of the business that the effort and cost are worthwhile. Once companies start investing in data analysis, it's hard to stop because they find that data-analysis results are far more important to the business than the cost. These projects are ultimately self-sufficient in funding. In some cases, companies start with small data, and after discovering that data analysis brings important insights, they begin to invest more to consolidate larger datasets and conduct more advanced analysis. For other businesses, small data can already meet their needs.

In either case, the business can benefit from data analysis: executives can see where the business can go in order to eventually improve their competitiveness, or, as the Marine Corps appreciates, understand the potential adversaries based on intelligence. To be sure, the value of data analysis is difficult to measure with simple numbers.

(Responsible editor: Mengyishan)