oCPC promotion effect is poor? Here’s the reason!

The cost of entering the second level has skyrocketed, traffic outside the business has entered, and search terms are confusing;

Conversion rates fluctuate, consumption continues to increase, conversion results are sometimes good and sometimes bad, and conversion rates are unstable;

Consumption is not going out, and there is no obvious increase in consumption display clicks. It is completely impossible to see that OCPC is running;

This is a situation we often encounter when running oCPC , so why does this happen?

If oCPC promotion is not effective, 80% of the reasons are due to the model.

Modeling is the soul of Ocpc. If there are problems in the accumulation of models in the first stage, it will be difficult to achieve good results in the second stage. It can be said that the basic factor that determines the quality of your oCPC is the model. Therefore, today we will focus on "How to model efficiently and stably" and talk to you about the timing of modeling, preparations before modeling, precautions for modeling, reasons for failure, etc.

1. Timing of model building

Building an oCPC delivery package does not mean that you decide to run oCPC tomorrow and then build an oCPC package and deliver it this afternoon. This requires us to make preparations in advance.

So when should we build the oCPC delivery package? This depends on the effectiveness of previous promotions, and we need to choose the appropriate time period and region to build our delivery package.

For example, if your performance on Saturdays and Sundays is very poor, with general leads, high costs and low conversion volumes, you insist on choosing Saturdays and Sundays to build packages. According to big data, the effect is actually not good. If your performance is good at night, you insist on building packages during the day to accumulate data. This is also unreasonable.

2. Preparation before building a package

Before building the package, it’s not that I want to run oCPC, I can just build the package directly. Before building a package, we need to organize the accounts first. Our account structure oCPC cannot be adjusted, so we need to adjust the structure manually first.

01. Creative Optimization We know that whether a bidding ad is clicked by users depends not only on its ranking but also on its creativity. Therefore, before building a delivery package, we must first optimize our creativity to at least make our creative click-through rate look acceptable. We should also highlight the selling points and direct benefits and describe them in the creative.

02. Keyword organization Keyword organization is very important. Many people have poor oCPC results because there is a problem with keyword organization, so the keywords must be organized.

3. How does oCPC achieve stable modeling?

01. Capture more accurate conversions

The data access method selects deep conversion behaviors, such as forms and three-sentence consultations. Do not choose shallow conversion behaviors, such as "clicking a button" or "consulting in one sentence", so that the model can be built more accurately.

02. Business settings are more precise

When establishing a plan, you need to choose a business. According to the characteristics of your own industry, you should choose the most suitable and most segmented industry business. Setting precise business for the plan can make the system allocate more relevant traffic to the plan.

03. More conversion data

During the data accumulation stage, try to accumulate as much conversion data as possible. The larger the conversion data base, the more accurate it is, and the better the effect of the subsequent intelligent delivery stage.

4. Modeling considerations

01. Pay attention to data verification. When building a package, we need to consider the time periods when our promotion is more effective, such as Saturdays and Sundays, evenings, and daytime. Choose the right time to check our data.

02. When binding the scope of targeted delivery, generally speaking, don’t throw all the plans into it, especially when some accounts promote multiple products, some products have a conversion cost of 150, and some products have a conversion cost of 50. It is obviously not reasonable to put them in one package at the same time.

Of course, sometimes we need to tie it together in order to advance to the second level, and we need to take it apart later.

5. What are the reasons for modeling failure?

01. Inaccurate conversion

For example, many travel agents use copying conversion buttons as a basis for adding followers. The data shows 10 followers, but the actual number of followers added is less than 10.

When running ocpc, it will collect the behavior of the conversion population, such as search terms, time periods, regions, and even historical searches and population attributes, which will lead to inaccurate conversion data.

02. Inaccurate industry

When we run the bidding ocpc, we must set the industry we are in and set the business scope of the plan as accurately as possible. Because once the conversion behavior occurs, it will be targeted at the industry.

03. Less accumulated conversion data

Generally, only when the conversion data reaches more than 15 in a single day or more than 90 in seven consecutive days can the conversion data be modeled. Of course, we cannot just pursue quantity here, quality is equally important.

Author: Houchang College

Source: Houchang College