This is how data people’s 2023 work plan should be written

How can a data person not know how to write a data table? 2023 is coming soon. Have data analysts thought about how to write their work plans for next year? If not, take a look at this article.

At the end of the year, it’s time to make a work plan for next year. So how should a data analyst write a work plan for 2023? You can’t just write “I pulled 500 tables last year, and I’ll continue pulling tables this year”! Moreover, the work plan involves performance appraisal. How can I write it to show my ability without digging a hole for myself?

1. How to write a regular work plan

For most routine work, you can use: GSA planning method

• Goal: Achieve XXX quantitative indicators this year
• Strategy: Achieve your goals through XXX methods
• Action: Divide into XX steps and complete in XX time

For example, for the sales department, you can write:

G: Achieve a small goal of 100 million

S: The old market maintains the current status of 80 million, and the new market achieves 20 million performance

A: Regularly visit customers in old markets every month, and complete 5 million tasks in new markets every quarter

The work plans of most business departments can be written in this way, but the G is different and the complexity of S varies. However, the work plan of the data department is not so simple because the nature of the work in the data department is different.

2. The particularity of data analysis

The special nature of the data department comes from the fact that it is a support department that has been given too high expectations. On the one hand, leaders all think that data analysis should play a high-level role such as "assisting decision-making", "empowering business", and "driving growth". On the other hand, the daily work of data analysis is largely consumed in temporary data collection. It is too busy to run data every day, so what high-level role can be expected?

This makes the data analysis department very entangled. If they only write "meet the business data collection needs", the leaders often express dissatisfaction. If they rashly write "increase performance by 5 million", they don't know where to start. Not to mention, similar dirty and tiring work such as data clarity, point management, and unified caliber, they work hard but are not understood.

Some other support departments also have similar problems, but not as serious as the data department. For example, the after-sales department, although it is also a support department, does not have high expectations. Therefore, the after-sales department only needs to calculate the corresponding service demand based on the business growth rate and arrange the labor force.

At most, you can add one to improve service experience/find secondary sales opportunities. For example, the IT department is also a support department, but its daily consumption is limited, and major development needs are often determined in advance by the business. According to the development needs, a development plan is formulated, and then it can be done step by step.

So, as a data department, how can we break the deadlock?

3. Start with target selection

In general, the data department has three major tasks:

1. Let the senior management see that the data department is working!!!
2. It has a certain promoting effect on business.
3. Complete a large amount of basic and routine data work.

Among these three, daily needs are the least of our worries. Business will always be important, and the key is how to efficiently solve needs and free up energy to do some valuable projects. As long as we report frequently, the top management will see our work, but how to make the top management believe that we have played a role in "assisting decision-making", "enabling business", and "driving growth" still depends on the implementation of specific projects. So the second one is the key to breaking the deadlock.

The second one is the most difficult to achieve among the three. Because if you want to be useful to the business, the cooperation of the business department is very important. If the business does not cooperate, no matter how hard you work on making models/reports/reports/dashboards, it will be useless. So the next step is to clarify: who is the best person to cooperate with?

4. Confirm the cooperation method

First of all, we need to distinguish the four types of business departments (as shown in the figure below). In principle, cooperation with elite teams is the best and can produce the most results. If you can't choose the second best, small teams are also good, at least they don't cause trouble. When cooperating with the other two types, you must be cautious in giving analysis results and suggestions, otherwise you will be criticized in minutes.

After looking at the target, you can further understand the needs. Some students will fall into the misunderstanding that the more complex the technology is, the better. In fact, it is not the case at all. In the eyes of the business, giving them something they don’t know is the most valuable. If it is already expected by them, even if you use a very complex model to calculate the result, they will say "I knew it a long time ago!" So when understanding the needs, it is best to feel the bottom of each department to see how much they know and what they want to know.

Common ones, such as:

1. New work in the business department. For example, new products, new promotions, and new activities. Any new work means that a complete set of data collection, data monitoring reports, and data review reports are required.

2. Classic business problems. For example, how to select successful products, how to measure the effectiveness of activities, and how to improve the efficiency of advertising. If the business wants to cooperate, these are suitable for in-depth thematic analysis and combined with AB testing to see the results.

3. More comprehensive data presentation. For example, if there were several tables scattered before, the business leader hopes to see everything in one table. At this time, it is very suitable to make a data dashboard based on BI.

4. Faster data response. For example, weekly reports were issued before, and business leaders hope to have daily progress updates (the sales department has the most similar needs). At this time, it is also suitable to make a data dashboard based on BI and issue it to each group.

There is another type of demand that is most often asked, which is forecasting! In fact, when the business is not sure what to do, they will ask: "Give me a forecast", but many questions are not forecasting questions at all. You must ask one more question:

• If the business says, "I don't know what the external environment is like, let me make a prediction," this is a data collection problem.
• If the business says: "What should I do to increase revenue based on the forecast?", this is a business optimization problem
• If the business says: "Predict sales so I can control inventory", this is also an optimization problem
• If the business asks: "Predict which product will be successful next", this is a test problem

In short, don't start using models when you hear about prediction. Instead, break down the business problems, collect data when necessary, diagnose business processes when necessary, and do tests when necessary. Output results from a perspective that is useful to the business, rather than just starting with a name.

Finally, after sorting out the needs of each department (as shown below), you can put them into the annual plan with some trade-offs, so that you can achieve your goals.

5. Clarify the implementation details

When implementing, please note that "productization" is the way out for the data department. Although it is just a number, it can be reported using BI/Excel/verbal, but try to use the system to achieve it. "At first, I was like a dog looking for numbers, and after reading the report, I thought the people were ugly" is the norm. Only by continuously implementing the system and making products can we ensure that the data department can continue to recruit people, constantly remind everyone of the value of data, and directly demonstrate the effectiveness of "digitalization" construction. Remember this.

Therefore, when clarifying the implementation details, BI development needs should be prioritized. All data monitoring, data evaluation, and data review needs that can be fixed should be developed and made into data dashboards.

All models, user portrait analysis results, and product classification results should all be entered into the data warehouse and CDP, and made into fixed update labels. Temporary and test activities can be formatted in a fixed format and automated reports can be used. In short, try to get rid of temporary data collection, so that the extra time and energy can be used to do more project work.

Finally, remember to confirm the delivery time nodes of each project. Generally speaking, it is better to deliver results in each of the four quarters of the year. At the end of the year, the business is rushing for performance, so the data output results are mainly data reports and analysis reports such as "boosting business" and "supporting big promotions".

In response, the beginning of the year is more suitable for doing some basic work, such as data integration, sorting, and fixed report development. The middle of the year is more suitable for coordinating business pilots, doing some complex tests/special analysis, and drawing some analytical conclusions to reflect the data analysis of "reducing costs and increasing efficiency" and "promoting growth".

The whole arrangement can be shown as follows:

The above is of course different for each company’s organizational structure, and the size and division of labor of the data department are different, so the specific circumstances may be different.

Author: Down-to-earth Teacher Chen

Source: WeChat public account "Down-to-earth Teacher Chen (ID: gh_abf29df6ada8)"