Advertising strategies for the Internet automobile industry

There are many common traffic allocation strategies in automotive industry advertising, including level market traffic allocation strategy, KA customer traffic inclination strategy, etc.

Question 1: When the company's KA customers (key customers) sign an annual framework/large order with us, the platform makes extremely high performance promises. If we only rely on the existing delivery logic, it is very likely that these promises will not be fulfilled within the specified period.

In the vast majority of Internet advertising companies, there are special policies to take care of KA customers. Common special policies include: exclusive product discounts, unconventional additional benefits, super high effect commitments, etc.

In order to win big orders, many small and medium-sized platforms will even make promises of results that exceed their actual traffic, and this situation is becoming more and more common in this era of sluggish growth in Internet users. In short, in order to take care of the feelings of KA customers, the marketing side will always try their best.

But it is obvious that KA customers are more concerned about whether the investment can really bring the promised effect, or it can be further understood as whether increasing the investment in this platform can really bring excess returns. If the platform wants to fulfill these promises, in addition to adding optimization measures at the media execution level, it also needs commercial advertising product managers to find ways to divert traffic.

As shown in the figure above, in the actual advertising process, KA customers placed large orders and significantly increased the daily budget. Without the support of traffic allocation strategies, they are likely to face the situation of unused budgets.

Of course, there is a prerequisite for this situation: the overall traffic of the platform is insufficient, the traffic utilization rate is already very high, and there is no budget for purchasing traffic from outside. Especially in vertical media, if customers know that the traffic is purchased from outside, they are likely to cancel the order directly. After all, the purpose of choosing vertical media for delivery is to target vertical media's precise and high-intent user groups. Instead of investing in external media through vertical media, it is better to invest on comprehensive media platforms.

In the absence of traffic support, all you can do is to consume the customer's budget by strengthening operational means, such as the most common methods of loosening targeting conditions and adjusting bids.

But it will also bring more problems. Although the relaxation of targeting will expand the traffic opening, it may lead to a decline in CTR and CVR because the targeting is no longer accurate. Adjusting bids can quickly consume budgets, but the overall ROI will also decrease, and the decline in these key outcome indicators will be more difficult to explain to customers.

At this point, it is time for the commercial advertising product manager to play his role. Although there is no way to create traffic "out of nothing", traffic allocation strategies can be used to bring certain accurate traffic to customers in a "fleecing" manner, ensuring that the budget is consumed without a significant decline in key outcome indicators and delivering the promised results.

1. Countermeasures: KA customer traffic tilt strategy

When the traffic volume is limited and the traffic utilization rate is already very high, it is not easy to find the precise traffic. There are usually two ways to think about it:

① Conventional thinking

Conduct deeper user behavior research and optimize algorithms such as look-alike to find more potential users. This method essentially follows the old path of expanding targeting, but the method of expanding targeting is more sophisticated than directly relaxing the targeting conditions in the delivery background. However, it is difficult to determine how much effect it will ultimately bring and when it will bring obvious results. Algorithm students need to continue to experiment, so we still need some "devious tricks" to save the day.

② Crooked ideas

Since the goal is to find more precise traffic targeting this car series, starting with the direct competitors of this car series is a very clever way. The target groups of several car series that are direct competitors will have many similar characteristics, and the possibility of achieving results by exploring within this range will be greatly increased.

This method is also easy to understand from a business perspective. When a user wants to buy a car, he or she will usually pay attention to 4-8 car series at the same time in the initial stage. Whoever shows themselves more to the users within this range will have more opportunities for users to understand them in depth, and the chance of a deal will be greater.

However, this approach of exploring direct competitors is not necessarily suitable for advertisements in other categories, because when a user wants to buy a car, his purchase range is already limited to cars. The result is either he buys car A, car B, or nothing. But it is rare that he goes beyond the category and ends up buying a motorcycle or a bicycle.

In the e-commerce sector, there are counterintuitive cases such as beer and diapers. That is to say, it is only interesting to study the problem of traffic distribution among competing products when the user's target category is limited. When the category is not fixed, conventional thinking is obviously more effective.

2. Implementation of the Strategy

The implementation of the conventional strategy is that the algorithm team develops a population expansion model for the target car series. After the model is trained, it will be pushed to the DMP. The media executor/agent can use this model on the DMP to generate a new group of population packages and push them to the targeting module of the advertising delivery platform for delivery. Finally, the delivery results are sent back to the algorithm side for subsequent model optimization.

The implementation of the crooked strategy is the same as the previously introduced strategy, which is achieved by increasing the weight of the target object in the reordering phase.

However, it should be noted that the KA customer car series weighting is not a universal strategy, that is, it is not a strategy that will be triggered in every bidding. The triggering condition requires that the target object and its competitors participate in the same bidding to trigger the strategy.

Secondly, before weighting the target object, the general strategy will be triggered first, such as: level market weighting, and then the special strategy will be triggered: KA customer car series weighting. Although both strategies are weighted, they cannot be mixed arbitrarily, otherwise the algorithm model behind them will become very confusing. Different strategies should be used to solve different problems as much as possible, and the strategies should be as independent of each other as possible.

3. Evaluation of Strategy Effectiveness

The effectiveness of conventional strategies can be fully evaluated through indicators.

First, how many new and accurate users can the new model find for us, and whether these users can support the consumption of the customer's daily budget. Secondly, how effective are these newly added population packages in their delivery? Will the CTR/CVR/ROI be lower than the average level of daily delivery? If it is only slightly lower than the daily level, such a result is acceptable.

The effectiveness evaluation of the unconventional strategy is more complicated. First, we need to verify how many new exposure opportunities we have actually provided to customers after weighting, whether these exposure opportunities have generated subsequent conversion behaviors, and the evaluation indicators of these conversion behaviors cannot be significantly lower than the daily delivery level.

Secondly, you should also pay attention to the account consumption of related competing products. If there is a sharp drop, there is no way to explain it. Normally, it is acceptable to control the traffic drop of each competing product within 10%. In actual applications, we may encounter a situation where the conversion efficiency of traffic snatched from competitors decreases. At this time, we need to re-examine whether the calculation logic of direct competitors is reasonable. If the calculation of direct competitors is unreasonable, all subsequent efforts to snatch traffic from competitors will definitely not be effective.

Finally, conventional strategies and unconventional strategies can be implemented simultaneously, so that the pressure on the effectiveness of the two strategies will be much smaller and the possibility of the strategies being effective will be higher.

4. Generalization of Strategy

There is actually no problem of generalization with conventional strategies, because this strategy is completely universal and can be designed using this approach when customers in almost every industry need traffic support. The crooked strategy is only applicable to highly competitive industries with a limited scope, such as automobiles, real estate, etc.

But the crooked tricks also bring us some inspiration. I believe that many product managers engaged in commercial advertising have encountered situations where they have to force a project to complete KPI quickly. In such a very urgent scenario, you can try to adopt the strategy of taking advantage of everyone to save the day.

However, it is important to note that strategies such as crooked tactics should only be used to deal with emergencies in the short term and should not be used on a large scale for a long time.

Author: Everything needs Jingsheng

Source: Everything needs Jingsheng