4 rules for operating advertising accounts!

Advertising is like getting married. The advertising plan is the bride and the advertising system is the groom.

Marriage is a matter between two people and requires both parties to be willing. If the bride is willing, open the door (set the budget and start the advertising); if the groom is willing, come and pick him up (place the advertisement). In this way, the account has a certain amount of volume and the marriage is completed.

The bride is the passive party, and the initiative lies with the groom. The marriage can only be concluded if he comes, so even if the bride opens the door, she may not be able to get married. But the bride can refuse, and she locks the door so no one can come in (shut down the plan).

The bride is reliable, just sitting there waiting to get married; but the groom is unpredictable, he may not come when he says he will, so the biggest feature of advertising is "maybe" - if you test a plan, the payment cost is very good and stable; if you can't test it, no matter how many plans you put forward, it will not work, which will make you anxious. The same goes for consumption. If you talk about volume, it can increase very quickly, with daily consumption increasing from 1,000 to 100,000 in just one night; but if you talk about no volume, there is no volume at all, even if you scratch the wall.

For example, this is the consumption data of a Toutiao account. It is like a roller coaster, up and down. It is as high as 130,000 and as low as 300, which makes the optimizer worry a lot.

For us, since advertising is uncertain, "waiting" has become an optimization idea for advertising, and it is a very commonly used optimization idea. Just do nothing, no need to adjust the bid or direction, just wait.

So what are you waiting for? Of course, wait for good luck to come... but you can't tell your boss that, ahem, you say, "Waiting is waiting for the plan to have more conversions and give the plan more learning time."

The specific application scenarios are:

Once you have set up the plan and started implementing it, all you have to do is wait and observe the data; if the plan doesn’t cost money, you generally just have to wait. The planning cost is high, but you can still wait...but is waiting useful? It is quite useful. You are not doing any work while waiting, but waiting for the system to do the work. If you adjust too much by yourself, there may be side effects and it is easy to mess up the plan. Therefore, novices just need to trust the system and wait for the system to take them flying. Experienced people can cut off points at necessary times to do some operations.

Maybe the advertisement will be put up - then I'll wait. Waiting can increase the probability of being flipped.

Because the advertising system is uncertain, we can only passively accept the arrangement. During the arrangement process, I found that "0 o'clock" seemed to be an opportunity for reincarnation.

We usually say "the cost is good today" and "the cost was high yesterday", as if we are allocated a certain amount each day. When the cost is high, we will also say "the cost is relatively high today, let's observe tomorrow", because there might be changes the next day.

Costs do vary from day to day. Let’s take a plan as an example and look at how its cost changes. This plan was just launched at 18:00 on the 8th. Apart from releasing the budget, no adjustments have been made in the past few days. A lot of money has been spent, which is quite representative.

Planned Operation Log

Let’s look at the data from the 8th to the 11th. It was found that its cost on the 8th was 30. Without any adjustment on the next day, the cost dropped from 30 to 26, and the quantity also increased. It can only be said that it was a good reincarnation.

Costs will change on their own from day to day unless you make adjustments. So when the cost is high, you can wait a day and see.

In addition to the cost, whether the plan can spend money after 0 o'clock will also be different. And it can probably be seen in the first 2 hours.

If more money is spent from 0:00 to 2:00, the account volume today will most likely increase; if less money is spent from 0:00 to 2:00, the account volume today will most likely decrease.

Still using this plan, we select the time-sharing data in the background's "Report" - "Plan Report", and then compare it with the previous day to see the trend of the data.

① The first two hours of the 9th consume more energy than the 8th, and the whole day's consumption is also more. Number 8 cost 80,000, and number 90,000 cost 480,000.

②The consumption in the first two hours of the 10th is less than that on the 9th, and the consumption throughout the day is also less. No. 9 cost 480,000 and No. 10 cost 430,000.

③The consumption in the first two hours of the 11th is less than that on the 10th, and the consumption throughout the day is also less. No. 10 cost 430,000 and No. 11 cost 250,000.

For new plans (the first three days of launch), it is easier to see explosive growth at 0 o'clock. I won’t give any examples here, just pay attention to it.

So what’s the point of saying so much?

No matter whether the cost is high or we plan not to spend money, we can wait a little longer, because there will be a chance of change after 0 o'clock. If the cost is high today, it may be much lower tomorrow; if we don't spend money today, we may spend money in a flash tomorrow. Generally, you can wait for 2 to 3 days.

The specific approach is: the cost is very high this year and we can’t bear it anymore, so we can suspend the period and wait and see tomorrow, maybe the cost will go down. Even if you don't plan to spend money, don't pause it. You can wait for 2 to 3 days, and there may still be a chance to spend money.

As for this system, although it is difficult to marry him, after marriage (investing a certain amount), he will still treat you well, and the headline cost is stable.

When we say maybe, we are mainly talking about consumption. The consumption is unstable, but at least it is acceptable. The worst case scenario is that the marriage does not take place. However, the cost is unstable, just like being hit hard after getting married, which will be a big trouble.

Now everyone uses conversion bidding, and the conversion cost is basically controlled by the bid. It is normal for the cost to fluctuate by about 20% relative to the bid. By "unstable" we mean "a plan that has accumulated a lot of conversions may suddenly have a cost that is much higher than the bid." When we run a media, the first thing we pay attention to is whether the cost is stable. The cost of Toutiao is basically stable.

How does knowing that costs are stable help us? You will have a better idea in your mind. What worries optimizers the most is the high cost. For example, the optimizer did not do anything, and the cost was 27 yesterday, but this morning he found that the cost was 30. This must have scared the optimizer, and he wondered what to do and whether to adjust the price. If you know that the cost is stable, you don't have to panic. You don't have to just look at the cost of a certain period of time, but can extend it to look at the whole day. Let’s look at a few examples.

I randomly selected 3 top plans and 2 days of data for each plan, which is a total of 6 days of data. In these 6 days, some people spent 4,000 yuan a day, and some people spent 400,000 yuan a day. The data difference is quite large, but it is still quite representative. Let’s look at the difference in morning and afternoon costs for each plan. (The data obtained in the morning is from 0:00-13:00; the data obtained in the afternoon is from 13:00-24:00. No adjustments were made to the bids throughout the day.)

You will find that the cost in the morning is always high, the cost in the afternoon is always low, and the total cost for the whole day is between the two.

We plot the data of these 6 days into a line graph, and plot 3 points for each day's data according to "morning, afternoon, and all-day data." We will find that it forms a series of V-shapes, some are big Vs and some are small Vs. This makes it more obvious that "the cost is highest in the morning, lowest in the evening, and the data for the whole day falls back to the middle of the two"

I randomly took 6 days of data and found the same pattern, which shows that "lower costs at night than in the morning" is a common phenomenon. So friends, if you have a plan with stable delivery, and you find that the cost is high in the morning, don’t panic, the cost will go down after lunch. The cost is lower at night. (The new plan is not like this. The plan for stable delivery is like this. The plan for stable delivery is like this.)

The cost of headlines is stable. Don't panic if the cost increases during a single period of time, as it can be reduced over the whole day.

Planned consumption for 3 days

So what does it mean to release the budget in time? Within 2 hours after crossing the finish line. It is definitely the safest thing to release the money when it is being spent quickly. The planned cost can continue to be spent and the fluctuation will not be too big. After this period of time, you may not be able to spend the money.

Advertising is like getting married. It is a matter between you and the system, and both parties have to be willing to get married. Although it is a matter between two people, the initiative is still in the hands of the other party. The other party is not reliable, so there is no guarantee that the advertising will be successful. When there is no other solution, "waiting" becomes a solution;

During the waiting process, "0 o'clock" seems to be a chance for reincarnation - even if no adjustments are made, the cost and magnitude of the plan may change. Therefore, if a plan has no volume and high cost, it is okay to wait for 2-3 days, and the plan may still have a chance;

It is uncertain whether you can get married, but once you get married, he will be pretty good to you - the cost of Toutiao is stable , and it is generally not the case that a plan that has accumulated a lot of conversions will suddenly have a cost that exceeds the bid by a lot. Therefore, don’t panic if the cost in a single period increases, you can extend it to the whole day. For example, if the cost in the morning is higher than yesterday, it may be able to decrease a little in the evening. (Some accounts have high costs at night and low costs in the morning. This requires the optimizer to observe it himself. The high and low costs of the time period are also related to the margin. For details about the margin, please see "Why is there a delay in the number of conversions")

We don’t know when good luck will come, but all we can do is cherish it once it comes. The flow is not stable and balanced, but it is bar-by-bar . When you catch this wave of water, quickly release your budget.