Baidu Search oCPC Optimization Guide

Many friends came to ask me, why did their search oCPC promotion suddenly lose its effect after a period of time? The volume was originally good, but suddenly it dropped? Does it have a decay period like information flow? How can we optimize this type of declining volume?

This problem is quite serious, so today I will do an analysis and help everyone find the direction of optimization. If you have a better idea, we can discuss it together.

At present, Baidu search OCPC will indeed have a certain period of decline. If the account has the following performance, it can be basically considered that the OCPC package is declining:

1) Judging from the promotion effect, the conversion rate may be gradually decreasing;

2) From the consumption point of view, it may be that the money cannot be spent, resulting in poor conversion effect;

3) There may be insufficient population coverage, because oCPC also expands the volume based on the keywords selected by the account. If the product demand cycle is longer, once the traffic is expanded, there may be a problem of insufficient traffic.

If you want to optimize, you still need to solve the problem from the root, and find ways to increase revenue: spend money and expand traffic; save money: reduce funnel loss as much as possible for incoming traffic.

1) When setting the initial CPA, it is recommended not to set it too high. It should be within a controllable range. For example, how many conversions are expected in a day, and then set the cost based on the existing budget.

If it is set too high, you will get fewer conversions within a limited budget. Over time, the system will roughly figure out the amount you need, and the traffic allocation will gradually become less and less. Therefore, it is necessary to set the CPA reasonably, and even set it slightly lower than the original expected CPA in the initial stage (not less than 80%), and then gradually increase it.

If the cost is high in the initial stage, it is not recommended to enter the second stage, as it is not conducive to gaining volume or reducing costs, and there is little room for subsequent optimization. Because the second order is modeled based on the first order data, if the first order data is not run well, the second order modeling will not be accurate enough. What the second order can remember is the cost required for the first order to achieve effective conversion, and the system will execute to expand according to this cost. If the cost is limited, the system will be constrained and it is possible that the oCPC package will not be able to increase in volume.

2) After steadily entering the second stage, pay attention to search term reports, optimization suggestion reports, etc., add business-related words in the reports, and continuously expand the traffic pool. Continuously adding potential conversion words to the account and providing the account with more conversion traffic is a positive operation. There is no need to worry about affecting the model. Doing it a small number of times will have little impact on the model, so duck does not have to be restrained.

1) eCPM traffic distribution is related to three factors: CPA, creative CTR, and landing page CVR. Therefore, in order to obtain more traffic and exposure, in addition to increasing CPA, creative CTR and landing page CVR are the key points we need to improve and optimize.

Improving CTR requires optimizing creative copy, style, and pictures. I will not elaborate on this here. You can go to the relevant courses in the special training camp.

oCPC conversion is divided into shallow conversion and deep conversion. When doing daily data, such as consultation cost, intention cost, and transaction cost, it is recommended that everyone do statistics and calculate the conversion efficiency. If the conversion efficiency is low, it is recommended to set the conversion to deep conversion to improve the conversion efficiency from the source.

2) The oCPC model is constructed based on the traffic running in the first stage of CPC. Whether the first-stage running volume is real and effective will affect the establishment of the model. Therefore, in the first stage, it is best to keep the efficiency at around 80%. The higher the efficiency, the higher the accuracy of the model. In this case, entering the second stage will have more advantages in expanding the second-stage model and can better slow down the decline period.

Author: Aichisem

Source: Aichi Junior Sister