The whole process of information flow data analysis, with examples~

I heard that everyone has been tortured by data analysis .

▲ What the hell is data analysis?

▲ Why do I feel like there are problems with this data everywhere?

▲ Oh my god, which data should I analyze first?

It’s painful, isn’t it?

Haha… However, today’s article may make you more painful.

In order to help everyone escape the obsession with data analysis, the editor today brings you a case article on "How to Analyze Data ".

The figure below shows the marketing data of an account. From your perspective, what do you think went wrong?

After the analysis, you can continue reading with your own answer.

Generally speaking, why do we conduct data analysis?

Reduce costs, increase conversations, increase traffic quality… and more.

But in fact, in the end we can all boil it down to one goal: to increase conversions .

Then we can start from this purpose when we analyze.

Now that the purpose is clear, which is to increase conversions, we can start with the results.

From the figure we can see that its clues are gradually increasing, but the clue cost has not decreased.

Well…from the analysis of the results, our customer acquisition cost is relatively high.

The cost of a lead is high either because our average price is high or because our conversation rate is low.

But judging from the conversation rate, its data is acceptable to us, indicating that the traffic quality is fine; the click-through rate has dropped slightly, and the average price remains high, so the conversation cost is also at a relatively high level.

Therefore, we can determine that the high cost of conversation leads to a problem of clue cost.

Once we have identified the problem, we need to break it down.

It is recommended that in cases like this, we can list a mind map on a draft or computer.

The cost of dialogue is high, and we can solve it from two aspects:

1. Reduce conversation costs
2. Increase the volume of conversations

• Reduce conversation costs

To reduce the cost of conversation, either lower the overall average click price to reduce costs, or increase the conversation rate to win by quantity.

Reduce the overall average click price: We can achieve this goal by filtering out words with high average click prices and low conversion rates.

Improve conversation rate: Conversation rate is often related to traffic quality and conversion guidance. Then we can find our own weak point in influencing the dialogue by analyzing the following four points.

Arrival Analysis

Load Analysis

Transformation Capacity Analysis

Traffic quality analysis

• Increase the volume of conversations

Increasing the volume of conversations is simply a matter of increasing the quality and quantity of traffic .

This requires us to filter out junk traffic while increasing the amount of traffic. Again, we can achieve this by segmenting words.

Our initial goal is to increase conversions, so we can first filter out words with better conversions and then classify them.

High average price and good conversion: add keywords first, expand the volume, and then optimize the creativity to control the traffic.

Low average price and good conversion rate: combine price increase and matching.

Once the plan is determined, we can execute it according to this optimization plan.

According to the above operations, we can basically divide it into three steps:

1. Reduce the overall average click price

2. Increase the conversation rate

3. Increase the volume of conversations

So, the question is, which step should we take first?

Is it 123 or 321, or 213, 231...

In data analysis, which one you operate on first or later can lead to huge changes.

For example: When we reduce prices first, it is possible that the money will not be spent, so we need to increase the volume first and then reduce it.

So, dear readers, what do you think about the above order of operation execution?

Source: Houchang College ( semlianmeng )