Why personalized service is so hard to implement
Statistically, 95 percent of companies are not using their data, and 39 percent of marketers believe they are not able to correctly anticipate their customers' needs through their data. With such a large amount of data being left vacant and a significant percentage of it always being wasted, it's not that personalized services are difficult to implement, but that the data simply isn't being used as effectively as it could be. At the same time, it is also limited by the current level of technology, data is difficult to transform into services. It's as if you have a piece of jade, but you just don't have the delicate handcrafting skills to carve it into a valuable work of art.
The white paper, "Analytics: Big Data in the Real World," released on March 12, 2013, provides five key recommendations for the application of big data, including "putting the customer at the center" and developing a "big data strategy plan" up front. Develop a comprehensive and complete enterprise "big data blueprint"; start with existing data, set and complete short-term and phased "big data strategic goals"; according to business priorities, gradually establish an analytics system, and gradually improve the "big data analytics capabilities". "Big Data analytics capabilities"; customize measurable metrics to analyze "Big Data ROI (Return on Investment)".
The conclusion comes from a big data study conducted by IBM and the University of Oxford***. The project surveyed 1,144 business people and IT professionals in 95 countries and 26 industries around the world, interviewing more than 20 academics, business subject matter experts and corporate executives.
The ideal personalized service
To provide the ideal personalized service to users, companies must master two things: first, how to fully understand the user's personality through data; and second, to reasonably control and design the personality of the service.
Understanding user personality means providing users with the products and services they want. First of all, the enterprise needs to be in a huge database, to find out the most gold data; secondly, the data performance of the same user is divided into a class, based on user data performance design targeted services. Here, the key to whether the enterprise's services can be done is whether or not to seize the most core data. But have to think about a problem is: through the data analysis of the categorization of too many services, whether it will lead to increased management costs, while reducing the efficiency of the service?
Personalized decentralized units can be large or small, as large as a group of customers with the same needs, and as small as each user is a personalized demand unit. And too much dispersion of personalized services will increase the cost of service and management complexity of the enterprise, so it is necessary to reasonably control and design personalized services. Consider whether all the data provided should be converted into services? Are the increased costs proportional to the actual benefits? What's the point if the increased cost of the service doesn't translate into a better return?
In short, the biggest difficulty for enterprises to realize personalized services is the reliability of key data and the manageability of management costs. Specifically, the starting point of personalized service design is the analysis of key data, if the data screening and analysis of errors, the results can be imagined; personalized service comes with a variety of cost increases, such as data management. Personalized services to some extent can only be consumer groups as a unit, rather than each consumer, and must take into account the actual cost of the enterprise investment and return on revenue.
Four steps to personalized service landing
1. Extract massive basic data. The enterprise has big data is like having a gold mine, the gold content of this gold mine is high or low, directly affecting how much gold can be extracted. Similarly, the quality of big data is good or bad, also directly determines how much data the enterprise can subsequently utilize.
2. Mining useful core data. Refining useful data from basic data to organize and match is data mining. Data mining requires a professional data company to operate, and it is difficult for the average enterprise to have such expertise. So, are enterprises willing to open up their core data? Do they have the financial ability to hire a professional company? All these need to be weighed.
3. Response to marketing data. After the data results are used for marketing, companies have to respond. After the data is mined and can be applied to a certain market segment, the company also has to develop a targeted marketing strategy.
4. Maintain member service data. The execution and implementation of marketing programs and follow-up services to further test the management and resilience of enterprises.