Are those extremely low-level information flow ads really effective?

The online game advertisement mentioned by the topic owner has a style like this. Let’s take a look…

One respondent answered this question from another perspective. I believe it also solved many of our doubts and can give us new thinking in marketing .

The answer is as follows (answerer: Su Su from HULOO):

I will answer your question from another perspective, thanks to an article I read before ( Guokr.com : Why do fraudulent text messages look so stupid? Scammers are not stupid, fools are really stupid).

We always receive some scam text messages, such as "Congratulations, you have won xx prize, log in to xxx website to claim it" and so on. Such a stupid text message, it's obviously a scam.

So here comes the question. Why are these scams still being used even though they have been exposed thousands of times? Why are scam text messages so stupid? Are the scammers stupid? In fact, scammers are not as simple as we think.

Starting from the "Nigerian Prince"

Anyone who has used email knows that spam is rampant, including some fraudulent emails. The most popular one is "Nigerian Prince". The content of the email is roughly as follows: A very senior Nigerian official wants to transfer a huge amount of money abroad in the form of "state secrets" and needs to borrow your name and bank account. After the transfer is successful, you will be given 10% of the huge amount of money (tens of millions of dollars) as a reward. If you agree to cooperate with these senior officials, after a period of time, the scammers will ask you to advance some "insignificant" handling fees and take care of the officials' expenses on the grounds that there are minor obstacles in the progress of things. And after you pay a few times, the other party disappears without a trace...

Such a ridiculous and absurd method has been passed down for decades, starting from the era of paper letters, through faxes and then to emails.

What the researchers want to figure out is why these emails, which are not well-written and can be received with a simple Google search, are still being used by scammers? Why don't they create some more realistic and original stories?

Set sensitivity based on loss

Before answering this question, let's turn our attention to fire alarms, which seem to have nothing to do with scam emails.

In the dormitory, the fire alarm often goes off in the middle of the night, but it turns out to be a false alarm. Why is it so unreliable? This is probably because the alarm itself has "limited capabilities". The sensor is always unable to distinguish between a real fire and normal conditions such as smoke produced by smoking, changes in air humidity, and high temperatures. As a result, two types of errors will occur: ringing when it should not (false positive) and not ringing when it should (false negative) . However, when one of the false positives and missed negatives becomes smaller, the other becomes larger . If the alarm is adjusted to a higher sensitivity, it may go off randomly when there is no fire due to dry air, smoking, high temperature, etc., but if there is a real fire, the chance of it failing will be much smaller; if the alarm is adjusted to a lower sensitivity, it will not go off randomly, but it will be more likely not to go off when there is a real fire.

The sensitivity of the alarm is a double-edged sword. The appropriate sensitivity depends on the loss (i.e. cost) caused by the two consequences of "false alarm" and "missed alarm". For an alarm, if it sounds randomly when there is no fire, the loss will not be great and at most it will make people feel bad. However, if it does not work when there is a real fire, it will cause huge property losses or even life-threatening danger. Weighing the pros and cons, we would rather adjust the fire alarm to be more sensitive. A false alarm is okay, but the terrible thing is that a disaster happens without us knowing it, resulting in damage to property or even life.

In fact, this problem also exists in many other scenarios, but the higher the sensitivity, the better. For example, in court, if the evidence against a suspect is insufficient and he is in a state of "half guilty and half innocent", the court will make two mistakes like a fire alarm: accusing a good person of being a bad person (false alarm) and letting a bad person go as if he were a good person (missed alarm). In this case, if the sensitivity is adjusted higher, it means "it is better to kill a thousand people by mistake than to let one go"; if the sensitivity is adjusted lower, it means "never wrongly accuse a good person." Comparing the losses caused by the two situations, if a good person is wrongly convicted, even if there are only a few such unjust cases, the credibility of the court will be greatly reduced, which is much greater than the loss of temporarily letting a bad guy go. Therefore, most countries will set a lower sensitivity here, which is the so-called "innocent until proven guilty."

The liar is not as stupid as he looks.

Back to the original question, what is the relationship between sensitivity and scam text messages? Microsoft researchers believe the scammers' emails have something in common with the above scenarios.

The goal of the scammer is to find a few victims among the vast crowd and defraud them of their money. To use the terminology we are used to, if a scammer has a KPI, his KPI is based on "transaction amount" rather than "number of consultations". After all, the fraud process is not just about sending an email. The key link lies in the continuous communication between the scammer and the person who has taken the bait. Only when the person has finally transferred the money can the scammer complete an "order."

The original email was a way to screen people from the vast sea of ​​people. Anyone who does SEM knows that the reason we set so many keywords and add so many negative words is to attract precise audiences. Only when potential customers enter the consultation stage can the conversion rate be high. The same goes for scammers. Their target audience is people who lack social experience, are not vigilant enough, and have money.

At this point, the quality of the fraudulent email is equivalent to the keywords set by the scammers, which will first screen out a part of the people who may eventually send money. This is the “sensitivity” mentioned above. For scammers, there are two situations for sensitivity settings: if the email quality is too good, it is easy to give false positives: people who are not easily deceived in the crowd are lured out and used for fraud; and if the email quality is too low, it is easy to miss negatives: stupid and naive people in the crowd are missed and not deceived.

From the perspective of the scammer, if the email is too vivid and interesting, it will attract a large number of "hookers", and the scammer will then carry out the next scam steps one-on-one, and the cost to the scammer will be very high. If there are too many people to contact, more "customer service" will need to be hired, more bank accounts will need to be opened, and more false materials will need to be created. In the end, these people, who are not stupid to begin with, will discover the truth after in-depth communication and refuse to pay. The scammer's "conversion rate" will be very low, and all their efforts will be in vain. The fact that this old-fashioned scam email can still attract people to take the bait shows that those who take the bait are easy to deceive.

To sum up, the probability of completely defrauding a person of their money is low, the cost of a false alarm is not small, and the cost of an omission is not high. These three factors combined will force scammers to reduce the sensitivity of their phishing. They deliberately send some very fake emails so that the victims who fall into the trap are "really stupid" ones, which can reduce the risk of contacting the "fish" and not getting a penny in the end . On the one hand, this shows that the cost for scammers to commit fraud is getting higher and higher nowadays, and this industry is becoming increasingly depressed. On the other hand, it also shows that the scammers are actually very business-minded.

Back to the Youku issue itself

Game ads are used to screen users. If you don’t like it, then naturally you are not its user. If you make an advertisement look impressive and attract a lot of clicks, but it turns out to be a low-level game , this will waste a lot of marketing costs (because many ads are pay-per-click ).

So, if you are not the customer base, there will naturally be someone else who is. Don’t expand your positioning to the entire society. Existence is reasonable. People who play this type of game will naturally be interested in this type of advertisement.