How to avoid invalid traffic and improve traffic quality in bidding promotion?

Today’s sharing content is about the details of display - traffic quality .

1. Concept

Anyone who does bidding knows the marketing funnel. The more impressions you have, the more chances you have of getting more clicks, thus more transactions.

However, simply pursuing performance improvement may not necessarily lead to the expected funnel, and it is likely to result in high consumption and poor conversion.

Therefore, we need to be very detailed in the display step, instead of blindly raising prices and matching. Where to start specifically will be shared in the operation section below. First, let’s understand the composition of the display together.

I roughly divide it into three parts: efficient display, ineffective display and inefficient display.

▌Efficient display:

It means that the customer actively searches for a keyword and is matched with the advertisement we place for this keyword. This kind of display is what we SEMers want. It not only has a high conversion rate, but also allows us to control it well based on customer needs and our own budget.

▌Invalid display:

For example, the display of malicious clicks from our peers is ineffective for us. Please refer to the previous article and reduce this part of the consumption by denying the IP.

▌Inefficient display:

There are several situations. The most uncomfortable one is explained by a picture. Baidu traffic distribution is as follows:

The first picture is a news page (PC version) that I was redirected to after clicking on a pop-up ad. When I was browsing the news, ads similar to information flow appeared below, so I clicked on the two small ads for shared office and mobile phone.

The second picture shows an ad triggered by me clicking on a shared office, which triggered a precise keyword search, even though I did not actively search.

The third picture shows me clicking on an ad on my phone, which triggered a precise keyword search, but I actually didn’t actively search.

In this case, the keywords seem very accurate, and they will not be denied if you look at the search term report. This type of customer may only visit your website once, and it is difficult for you to block the IP and visitor code, but most of them are invalid customers and have no intention. The point is that this is unavoidable. If you want to advertise, you have to spend the money, there is no room for negotiation.

This is considered a type of inefficient display. There are other situations that can also be considered inefficient displays. For example, the search for keyword A triggers your keyword B , but different creative ideas lead to different click-through rates, and the relevance of the landing page that comes in may also be worse. In the long run, the conversion rate may be lower.

There are still many situations waiting for your exploration. For example, our keyword report shows 20,000, but you may see that the search term report shows only 18,000. Where did the missing display in the middle go?

2. Operation

After we knew this display concept, I quickly looked at the real-time visitors and found that there were indeed many of these seemingly accurate words that clicked in but bounced out immediately. This may be one of the reasons why I have not been able to reduce my bounce rate.

However, we SEMers cannot prevent the rules of search engine traffic distribution, so we should first do the part we can control when optimizing. Here is my personal set of operation ideas for your reference:

1. Comparative keyword display

Open your own key keyword table. If you haven’t had time to compile this table, just open the account or client, sort by display, find the keywords with higher display, and see which ones we want to promote the most.

Then compare the keyword display in the past, check the keywords with obvious changes, and start to investigate the reasons for the display changes (matching mode, price, market environment, etc.);

For example: Keyword A, today's display is 20,000, and the previous search term reports are all around 10,000; if no operations have been performed on A, then the main influencing factors may be increased user demand or the bankruptcy of one of your peers, resulting in an increase in your traffic; if there have been operations such as raising prices or relaxing matching that have led to an increase in display, then look at the second step to analyze traffic quality.

2. Analyze traffic quality

The front-end panel data looks at search terms and IP access depth, and the direct effect looks at conversion.

For example: Still using keyword A, we want to see what the traffic quality will be like after its display is improved! The most direct way is to look at the conversion rate. If the conversion rate improves, then that’s great.

So what else can we look at besides conversions? We can also check the search term report to see if the keywords it matches are the ones we want, and then look at the IP access track and depth corresponding to this keyword in real-time visitors.

If everyone comes to view the news page, they may have encountered an SEO from the same industry, in which case we will just operate on him without explanation.

3. Through the second step of analysis, we found that the traffic is of high quality. We can continue to increase the volume or observe and decide according to the budget

For example: Still using keyword A, we found that it has converted very well recently, so what should we do next? Maybe many people are ready to operate it and bid up the price! Here is the key point, pay attention - directly raising the price to match the price cannot be defined as wrong, but it may not be the best way, it is just simple and crude!

What if conversion rate drops after adjustment? Therefore, it is recommended that you add it through the search terms of keyword A, expand the keywords of A at the same time, and then add the new keywords with a bid slightly lower than A, while ensuring that it does not affect the existing effect of A.

4. Through the second step of analysis, we found that the traffic is not very good, so we can consider it from the following aspects (this example will be a bit long)

For example: Still using keyword A, I found that there was no difference in conversion after the display increased, and consumption increased! Let’s eliminate the reasons why increasing impressions did not increase conversions?

We can filter keyword A in the search term report and view all matching search terms, which will result in several situations.

One is a word we don’t do! For example, we find that the words matched by keyword A are not the products we want, or our products are not the words we want, such as some cost words.

If we don’t use cost words, we will first negate the unwanted words, then find out the root words of these words, such as how much, price, etc., and negate their phrases into the unit to avoid other invalid words from triggering;

Second, the matching is chaotic! We found that many search terms are our existing keywords. Let’s take the cost word as an example. If we promote the cost word, we will set up a special unit to write highly relevant creatives and highly relevant landing pages, and then negate the units that are easily triggered by mistake, so that they will not match keyword A.

For example, when a customer searches for my keyword "How much is A", the keyword "A" is triggered. The search term report shows that this word has been added. Then we can just remove "How much is A" in the unit where A is located.

Third, the words that are matched are very precise and are not our existing keywords. In this case, we can check the IPs of these search terms to determine whether they are from the same industry. If they are from the same industry, we can refine the IPs. Newbies can refer to my previous article. If the IPs jump within seconds, it may be the traffic distribution of Baidu. Consider reducing the matching and price to reduce some invalid impressions.

What if after completing all the operations, the results are still poor? That may be other issues such as customer service or page conversion strength, which need to be analyzed comprehensively.

5. The treatment of display decline is similar

The display of keyword A suddenly decreases. We first compare it with our own operations. If no operations are performed, it may be that our peers have made adjustments or new companies have joined the bidding.

If you have performed operations yourself (such as lowering prices), compare your operations to make a judgment. The final destination still points to the quality of keyword traffic! If the traffic is good, seize the good traffic and expand keywords; if the traffic is poor, optimize it.

6. Words that have not been displayed for a long time

These are usually unpopular words or customer search words that you have added, and you can consider displaying them first at a wide range and low price.

3. Mentality

1. Due to the particularity of the bidding industry, you must see the account data to make adjustments based on it, so my ideas will have many loopholes when compared to everyone's accounts. The experts and novices can only use it as a reference, and never treat it as your own operation. It must be based on the actual situation of the account.

2. The conclusions you draw through testing are your own. Even if others say they are correct, you can remain skeptical and test them yourself.

3. There are many keywords in the account, and it is impossible to perform such detailed operations on each word, so we can select key words for observation according to the 2/8 rule.

4. Learn how to use Excel in SEM , which will save time and be more effective.

5. This is the most important point: the account is the second partner of SEM , so you should be gentle and careful with it! After all, everyone can be happy only when the wife is happy, and the account is no exception. I wish you all happiness every day.

The above is the sharing content of the author [Wenhao]. If you have any questions or any ideas, you can leave a message below the article to communicate and we can improve together. I am your little bidding brother. The next article will discuss data analysis. See you there.

Source: Wenhao