Building a data operation system: from a “data machine” to a “growth engine”

In the digital age, data operations have become a key driver of corporate growth. However, when building a data operations system, many companies often fall into the dilemma of "running data machines" and fail to fully realize the value of data. This article will explore the essence of data operations and analyze how to transform from a simple listing of data indicators to a set of operating mechanisms that can drive overall business growth. Through steps such as consensus goals, task decomposition, and monitoring execution, companies can transform data operations from backend support to a true "growth engine."

Many students don’t understand the difference between data operations and data analysis. They are confused when they hear about “building a data operations system” or “establishing a data operations mechanism”:

• What’s the difference between this and making a set of data indicators?
• Why did the operations department ignore me even though I set the data indicators?
• It seems that every type of operation has its own indicator system, so how can we operate with data?

Today we will give a systematic answer.

Let’s first ask the most critical question: For data operations, is the focus on data or operations?

data

operations

one

screen

screen

think

Test

hour

between

01 The relationship between operations and data

Note that operations is a big job with many branches. If you look at each type of operations separately, it has its own set of indicators.

This is the essence of the problem mentioned at the beginning: "Operation already has its own indicators, but they don't look at the indicators I set." (As shown in the figure below)

Because in essence, operations are a supplementary job. In theory, if the product is strong enough and the product is invincible, then there is no need for operations. Users will rush to buy out-of-stock products and have so much fun that they don’t want to leave. What’s the point of operations?

But most products and commodities are not so powerful, so they need operations to assist, through user incentives, promotional activities, content dissemination, commodity operation and other means to keep users fresh and promote users to continue to be active and pay. As the saying goes: "If the product is not enough, the operation will make up for it" means this.

Therefore, operations personnel will pay special attention to data, and especially to the "natural growth rate" in the data - all their work is superimposed on the "natural growth rate".

• User operation - the natural user conversion rate is 5%, which is increased to 10% through incentives.
• Event operation - natural sales volume was 10 million, and the event increased it to 20 million.
• Product operation: The remaining inventory is expected to be 18 weeks, and it will be cleared through 9 weeks of operation.
• Content operation - naturally read 10,000 words, and write a million-level hit.

And so on

When operations demonstrate their achievements, they will always bring data. Therefore, each department has often established its own data indicators.

02 The key to solving data operations problems

If data operations focus on data, they will eventually become data runners. Operations already have data indicators, so you just have to wait for them to rush you to get data, and they do it urgently. There is no need to set up a separate position for this. Just spend 6,000 yuan a month to hire a walking SQL machine and let the data department prepare a large wide table.

In fact, many companies do this, which is why many students have doubts at the beginning. Today we will not talk about these low B companies, but share the real goal of data operations and what big companies do with it.

It seems scientific for each operator to look at the data separately, but it actually carries original sin: the operation work itself requires mutual cooperation, but the interests of small groups in various departments are naturally in conflict.

• User operations hope to attract people by distributing coupons, and the profits of product operations are directly squeezed out.
• Event operators hope to get the greatest short-term benefits possible, and the pace of user operations is directly affected.
• Product operations hope not to disturb hot-selling products and to have more leftover goods, but both activities and user operations require the support of hard goods.
• Content operators wrote articles such as "Shock! Jack Ma! Huawei! Boiling!" and each one was read by millions, but in the end there was no conversion.

In a job that requires mutual cooperation, if each department evaluates its own data, it will inevitably lead to mutual sabotage between departments. Therefore, the better use of data operations is not to use it as a running machine, but to establish a data evaluation mechanism based on the overall goal, so that all departments can jump out of their own small circles and serve the overall interests. This is the original intention of establishing this position and the real value of the position.

So the answer to the question at the beginning is B. The essence of data operation is operation. It is to establish a set of assessment mechanisms that guide the work of each sub-operation team based on the overall goal. It is essentially a working mechanism, so it requires consensus goals among departments and coordination and cooperation to replace the state of fighting alone, so that it can play a role.

03 Methods for building a data operation system

Step 1: Agree on the overall goal and formulate overall tactics

Each operation team agrees on the overall annual goals of the major departments (such as DAU, conversion rate, sales amount, etc.), and chooses tactics to achieve the major goals, breaking down the major goals into each month. Note: The decomposition method is not necessarily based on monthly averages or past trends. The decomposition method may be related to the tactical selection (as shown below)

Step 2: Set phased priorities and assign tasks to each group

Step 3: Decompose the phased indicators, and each department will implement, monitor and provide feedback

This step is the monitoring process of regular operational data indicators, which will not be elaborated in detail.

With the first two steps, the operational work at each stage has a clear main task, so there is no need to worry about "why the short-term activity rate has dropped", "how much should be written for natural growth", "how many customers have complained fiercely" - as long as the overall goal is achieved. For details, each team can find improvement points when reviewing themselves.

Step 4: Monitor the progress of implementation and review the results from small to large

At this time, we must keep in mind the three principles of review:

1. Don’t question strategy until it’s executed
2. Do not revise strategy when inputs are adjustable
3. Don’t question the direction until the strategy fails

When all departments reach consensus on goals, track progress, report problems, and work together at regular department meetings, the mechanism is considered to be functioning properly. This ensures that the overall goal is achieved to the maximum extent possible and reminds each team of its key tasks - so that they don't get overwhelmed by the small things happening in their own area.