The game of advertising space allocation in search bidding promotion

This article talks about the ad ranking of search bidding promotion. Interested readers can take a look.

As for the ranking of search promotion ads, practitioners in the industry know something about it, although everyone's understanding varies, and at different stages. I am the same. I dare not say that I have fully understood the ranking of ad positions. I can only say that I hope to communicate and share with you my current thoughts on the ranking of ad positions without reservation.

01

The most basic understanding of ad ranking is to understand it literally.

Search promotion is also called search bidding promotion. The best ranking is awarded to the one with the highest bid. Whoever is willing to pay the most for this ad space will get this ad space. This logic is simple and crude, and it seems very fair to everyone, and advertisers also accept this rule.

Simple things are bound to have loopholes. When ranking is calculated only by bids, advertisers can bid a little higher or much higher than the next person, and the final ranking will still be the same.

For example, there are three advertisers, let's call them A, B, and C. They bid 10 yuan, 5 yuan, and 2 yuan for the same keyword, respectively. Then, among the three ad positions, A ranks first, B ranks second, and C ranks third. There is nothing wrong with this ranking. If B bids 2.01 yuan, it can still be ranked second. If the fee is deducted only according to the bid, it is highly likely that everyone will not actively bid high, which will eventually affect the platform's revenue.

This problem is easy to solve. Almost all platforms now use a two-price auction billing method. The final fee is calculated based on the bid of the next advertiser plus a very small amount, usually 0.01 yuan.

This setting has been praised by everyone as Baidu's conscience. Previously, A paid 10 yuan for one click, but now it only costs 5.01 yuan, which is nearly half the cost. In fact, if A is not a fool, even if the rules are not changed, he will find out that 5.01 yuan can rank first through continuous testing, but in the end, one click still costs this much.

02

The bid ranking combined with the generalized second-price auction method seems perfect, but loopholes will be discovered over time.

There is a kind of company with a lot of money but no professional optimization team. After opening an account, they just bid very high, because the final billing will be based on the next one anyway. They just write whatever advertising ideas they want. For them, the copywriting level is so high, and they only strive for the best ranking, so as to ensure the best promotion effect within their ability.

For Baidu, it is different. If A bids 10 yuan and the ad has a 3% click-through rate, while B bids 5 yuan and has an 8% click-through rate, and C bids 2 yuan and has a 6% click-through rate. Then for the same 100 impressions, A earns 15.03 yuan and B earns 16.08 yuan. In comparison, B is the big sponsor, so why is the big sponsor ranked second?

Therefore, the most cost-effective way is to sort by ad quality. In the current Baidu official ad ranking calculation rules, it is stated as follows: Ad ranking is based on the value of the ad, and ad ranking competitiveness = ad quality * bid.

In the above example, suppose the quality of A, B, and C is 3, 8, and 6 respectively. After the rule is changed, the ranking competitiveness of ads is 30, 40, and 12 respectively. Then the ranking of the top three ads can be changed to B, A, and C.

B is ahead of A. A's bid is 10 yuan, so the charge to B will not be 10.01 yuan. Therefore, the charging rules also change accordingly, and a quality ratio needs to be multiplied. B's charge becomes: 10*3/8+0.01=3.76 yuan.

This will not exceed B's bid of 5 yuan, and it can tell advertisers that improving quality can reduce the cost per click, and everyone is very satisfied.

In fact, B may be the second largest player in the industry and just wants to compete for the second-place ad position. He optimized the quality score in order to lower the CPC while still being the second-place ad position.

If B was in the second place, the deduction fee would only be 2*6/8+0.01=1.51 yuan. Even if the rules were not changed and the quality score was not included in the calculation, it would only be 2.01 yuan. Now it has been inexplicably sent to the first place, and it needs to spend more money.

From this perspective, the platform is the biggest beneficiary of such sorting rules.

03

Capital is greedy. From the perspective of the platform wanting to make more money, is there anything that can be improved based on the current sorting rules?

Some time ago, I got a patent from Baidu on advertising ranking (follow the official account "Dewu Notes" and reply "ranking patent" to download the full text of the patent), and I found a lot of interesting things in it.

The background of the patent states that the current technology cannot allocate ad space and pricing globally, resulting in low overall advertising yields. In other words, ranking and pricing technology can still be optimized to make more money.

What exactly is the problem now? It can also be seen from the core changes in the patent.

The basis for ranking in the patent is expressed as overall exposure value, which is actually the same as the advertising ranking competitiveness mentioned above. The key lies in how this is defined.

The inventor of this Baidu patent believes that different ad positions naturally have different click-through rate attributes. At the beginning, the click-through rate of ads in high-quality ad positions will naturally not be bad. However, excellent advertising creativity may not have a significant increase in click-through rate because it is in a relatively backward ad position.

If we can reasonably evaluate the value of excellent advertisements, we may be able to achieve better click-through rates and create higher value.

The method proposed by the inventor is to obtain the bid of each advertiser and the click-through rate of each ad position under the premise that all ad positions are available. The bid is multiplied by each click-through rate and then added up to calculate the overall exposure value. After sorting, the first place gets the first ad position.

The next allocation is not in order, but the first ad slot is left out. Among the remaining ads, the bids are multiplied by the click-through rates of the remaining ad slots and added together. The first one after sorting gets the second ad slot, and so on.

Such changes, on the positive side, will mean that excellent ideas will have more opportunities to be discovered. On the negative side, the competition will become more intense, Baidu will make more money, and we will have to spend more money.

04

In fact, it is not so easy for us to spend more money. After all, no matter how it changes, the deduction will not exceed the bid. It is fine to control the bid well. It is also very nice to have a better ranking.

The extra money Baidu earns is from those who bid at random. They think the deduction is 0.01 for the next person, and it doesn't matter how much they bid. In fact, Baidu is trying every means to get closer to their own bid.

Price is always a part of the game between advertisers and platforms. Under the same effect, advertisers want to spend less money, and platforms want to make more money. It is indeed not easy to find a balance point that everyone can accept.

Putting aside the price part, let’s take a look at the search ad rankings.

According to Baidu’s official statement, ranking is also affected by quality, that is, click-through rate, creative relevance, and landing page experience. According to the above Baidu internal inventor, it is affected by the click-through rate of each ad position.

It can be seen that the click-through rate is the indicator that the platform pays the most attention to besides the bid. I personally think that the two additional parts of the quality index, creative relevance and landing page experience, are actually linked to the click-through rate. If the creativity is irrelevant, there will naturally be no good click-through rate, and if the experience is not good after clicking in, it will make users feel disgusted with the advertisement, so they will not click on the platform's advertisements, reducing the overall click-through rate.

To be honest, if you want to increase the click-through rate, it is actually very simple, just think of some attractive copywriting. But after doing this and getting a good ranking, is the advertising really successful?

From the platform's perspective, it was a success. But for us advertisers or optimizers, there is still a lot to consider.

For the platform, click-through rate is conversion rate, and each click generates revenue. After we get clicks, the conversion rate depends largely on the quality of the people who click in.

What we need to do more is actually screening, especially in some relatively broad words. It is unnecessary to attract a bunch of people who are not very relevant to the product in order to increase the click-through rate. If you can successfully screen out people who are willing to buy your product, then it is a success, of course, this requires a certain amount of skill.

In my work, I have learned something very profound: high ranking does not necessarily mean high conversion, nor does high click-through rate mean high conversion. If high ranking means high conversion, then you can get the best ranking by bidding against others. In the end, ROI is the key. Under the premise of ensuring high conversion, the click cost is really nothing compared to the order amount.

It’s not that we don’t care about rankings. The official statement clearly states that different ad positions have different click-through rates. No matter how good your own creative ideas are, they still need the support of natural traffic.

When two swords are combined, they will be more powerful.

Author: Xu Dewu