2017年吹用户画像
2018年吹人工智能
2019年吹数据中台
2020年吹……
Answer: CDP! (Customer data platform customer data platform)
If you are really an old man who entered the industry in 2010 and hear this new word pop up, it must be a tight chrysanthemum. Because there are too many people who have fallen down on these new concepts in recent years. It can even sum up the five steps of new concept development:
1. Call out a new concept somewhere (80% probability is Ali/former Ali)
2. Moments of friends started to post articles and hype
3. Other companies started to recruit and recruited
4. The new leader wrote a bunch of people who could not understand their names ppt, start!
5. After half a year, get out of the roar of "You made a few dollars for the company!"
Let alone the “data center” last year, a bunch of traditional companies that blindly believe in digital transformation were killed, and the tragedy is still in sight. This time it is also called "Customer Data Center". It can be said that Mr. Chen was also skeptical and wait-and-see at the beginning. Until now, after pushing a bunch of projects to go online, I suddenly discovered: This may really be a new opportunity to be a data man! Because it overcomes the four major weaknesses of traditional data projects. Don't dare to monopolize it, share it with all the students
1. The four major weaknesses of traditional data projects
1. The root cause lies in data collection, not data analysis.
The root cause of all data items is data collection and data quality. However, data quality is precisely the link where enterprises are the most lagging, least willing to invest, and least able to see output. As a result, data projects often fall into the dilemma of having no rice in the pot, and they are all dealing with transaction data, and there is no other fart.
2. Myths and legends are rampant, and leadership expectations are too high.
Simultaneously with the thin investment, the expectations are too high. It is the leaders who expect to have a model of the mighty and invincible general that can duang know the sky from the top, the earth from the bottom, and the air in the middle. The ending is naturally easy to hit the street.
3. The daily presence is thin, and expectations are getting higher and higher.
Subject to 1, 2 constraints, people who do data projects often write powerful ppts, such as "detailed insights through user portraits", "precise predictions through artificial intelligence," and so on. Then a fierce man plunges into data cleaning and data warehouse It is impossible to extricate itself from the muddy ground of construction and access to data. Containing a ppt for half a year. As a result, in the past six months, the leaders' expectations are getting higher and higher. At the last look at the ppt, they shouted: "I knew it! What's the use of you doing this!"
4. Being out of touch with business, it is difficult to prove innocence.
Even if there is really a Shenwei Invincible General Model, it is 100% accurate to predict that a customer will buy the product, sofucking what! In traditional companies, this credit is still counted on sales. In Internet companies, this credit is still counted on commodity operations. Data cannot leave the business to independently generate revenue. It is also difficult to prove: Our company invested 5 million in Taiwan in the data , Earn back tens of thousands! In this way, when the performance is good, you can still be a mascot, and when the performance is not good, you can directly kill the sacrifice chess.
In the final analysis, most companies are not monopoly platforms like Tooteng's holding large amounts of cash. There is little data and it is difficult to realize, but I am looking forward to seeing results quickly. In this case, it is necessary for the data person to find a tool that can show up frequently, iterate quickly, and directly act on the business in order to meet the expectations of leaders. Similar to the past when doing user portraits, data middle-office, and artificial intelligence projects, the practice of picking up data and cleaning data for more than half a year is inevitable.
CDP has the opportunity to solve these problems.
Second, what is CDP
CDP is not a data system, but a business system. The basic logic is shown below:
The basic idea is:
1. Post tags to users based on data.
2. Screen target users based on user tags.
3. Develop an active strategy: at XX time, XX channels, push XX plans to XX users, and plan to achieve XX% response , XX users register/purchase/repurchase.
4. Develop a passive strategy: When an XX user has an XX action with our company on the XX channel, the XX rule is triggered to realize user registration/purchase/repurchase.
5. Select information channels, push strategy
6. Recover data, iterate tags
Is this set very familiar!
Students who have done CRM projects will shout: What is the difference between this TM and CRM!
Students who have done user portrait project will shout: What is the difference between this TM and user portrait!
The difference is only a little, but it has produced a very good change.
3. New opportunities for CDP VS CRM
Indeed, in terms of operational processes, CDP and CRM are exactly the same. But note: the core of traditional CRM is points and membership levels. First set the points policy, then set the bronze, iron, silver, gold, and diamond levels, and then perform actions based on these after it is operational. Points and membership levels are open, rigid, and stable policies that will not change in a few days, or even in a few years.
This leads to another problem: as far as data is concerned, the CRM system is less dependent on data and has less space for flexible operation. In actual operation, it often becomes small repairs and supplements every year according to the plan of the previous year. The existence of data is scarce, so it is difficult to reflect the value of data. For business, open, rigid, and stable policies are too rigid to flexibly deal with issues such as customer activity and conversion rates, and may even become a burden. For example, the huge stock of credits in the credit pool, whether clear or unclear, is a headache (as shown below).
The idea of CDP is much more flexible. It pushes marketing activities in real time, and you can try when you find opportunities. This can not only meet the needs of the business to respond to changes at any time, but also create more opportunities for data. Because a condition is good or not, it is essentially an experiment. As long as there are experiments, there must be early data insights, mid-term experimental design, and later effect evaluation. The presence of data is greatly improved, and business problems can be corrected at any time, and the degree of integration with business is also improved (as shown below).
Four, CDP VS user portraits to avoid backing up
Indeed, many user portrait projects are carried out according to the steps of: doing portrait-insight-grouping-strategy-effect, but there is a key problem: user portraits are often data-led projects, so it is difficult to get support on the business side!
The most common is that when you want to collect data, the business is 100 reluctant. If manual collection is required, they will decline and say: This reduces efficiency and the customer experience is not good; if system collection is required, they will decline and say: Business requirements are very tight and there is no time to collect data.
When you analyze the data, the business expects you to do something: it is very important to the company, and the business is 100% unknown, and the business knows how to land it at a glance, and it can see money when it is landed. There is one thing you do that the business already knows, and they jump out and spray: "I knew it! What's the use of what you do!"
90% of user portrait projects are so dead,
90% of AI-directed marketing, operations, and product projects are so dead
But CDP is different. It is a business system itself, and the first thing to transform is the way the business party works. Companies that have implemented CDP have transformed their business processes directly based on the logic of CDP and realized business process standardization (as shown in the figure below), so business participation is guaranteed.
At the same time, a discerning person will find that the core driving force of this process is the quality of data analysis. If there is a lack of in-depth data analysis, no matter how much experimentation, no matter how much pat on the head, it will fail. Even worse than without CDP. When there is no CDP, you can still give away: "Internal non-cooperation, external environment changes, although performance is less but satisfaction is improved"; with standard processes, whoever does not do well can be seen at a glance, and it is even harder to say To blame. At this time, more data analysis support is needed, and the importance of data is more obvious (as shown below).
V. Summary
For a long time, we have always described the best state of business and data as: sniper rifle-business shots, data mirrors. From blindly hitting with closed eyes to aiming and shooting is the result of the data.
From this point of view, CDP is a platform that can achieve this goal in the background of the company (the tool to achieve this goal in the foreground is called Sales Assistant, Sale Assistant or MCRM, but no one is hyping it). And this concept is still very popular this year, the bosses like fashionable things, so you can take the opportunity to make a good article.
Even if there is no platform or project, it is easy to achieve results if data analysis is done on conditional and event marketing. If you are interested, follow the down-to-earth teacher Chen. We will share a CDP actual case in the next article, so stay tuned.
Author: Chen grounded gas, micro-channel public number: down to earth school. A data analyst with ten years of experience has launched a series of data analysis courses and has more than 20,000 students.