How to deploy a huge amount of Qianchuan when cold starting a new account?

1. How to choose between ultra-fast delivery and professional delivery, and what are the differences?

Professional live broadcast teams must choose the Professional Plan. The Speed ​​Plan is a low-threshold paid purchase port designed for non-professionals.

The Speed ​​Plan is simple to operate and easy to place orders. Sometimes the ROI is also good, but it is unstable and there is no upper limit. You may get a ROI of 3 today and 1 tomorrow. It is basically impossible to keep the ROI stable in a range for a long time. The professional plan will be a little more complicated at first, but it is easy to operate and the ROI is stable and can be improved.

2. How to determine whether the Qianchuan account population is accurate?

Calculate the percentage of orders that are planned every day. If, in addition to the total order plan, the order ratio can be maintained above 40% for 5-7 consecutive days, then the account's population can be initially accurate, as mentioned in the previous article.

3. How to bid in the multi-product live broadcast room

If you have just started running and have not made any bids yet and have no idea about the bids, you can use the formula:

Bid = Live broadcast room customer price / expected ROI × 1.2

Assume: the average customer spending in the live broadcast room is 100 yuan, and the expected ROI is 3, then the bid is 40, and create a few more plans with step-by-step bids in the range of around 40.

4. When will the switch plan be implemented and what is the reference for the switch plan?

Looking at key data and closing plans can improve ROI, which will be discussed in the next article

5. Which data are most important for Qianchuan advertising?

Let’s talk about 5: bid, consumption, advertising ROI, number of advertising orders, and average advertising transaction amount

How to establish orientation

Optimize your goal here: there is only one option, which is the deal, and that is the only one you have in mind. As long as the first and ultimate goal is to close a deal, plans that are still building clicks and entries are all smoking in the opposite direction. Don't listen to those so-called teachers on TikTok who start by talking about two entry, two click, and two transaction plans that match each other. It's all bullshit.

The people who enter, click and make transactions are not in the same pool, so the established model deviates from the transaction model. The audience we need should be as precise as possible, and the first label must be transaction rather than clicks and views.

We must spend our paid money wisely. We only need a model of the people who make transactions. We don't want any of the people who look at the pictures but don't buy, or click on the pictures but don't buy. If we spend 1 yuan in the early stage to buy such people, we will have to spend 10 or 100 yuan to correct them later.

If there are too many plans for clicks and views, it will become a drag when we want to improve ROI in the later stage. The later the delivery, the higher the accuracy required, so don't deliver inaccurate amounts at the beginning.

The first is that if the price is too high and the plan fails, the loss will be minimized. The second is that if you directly max out the budget, there will be less opportunity to increase the budget to give the system positive feedback. You can increase the budget when about 50% of the budget has been spent, by a few hundred or a thousand at a time. There is no limit on the number of times you can increase the budget. After the budget is spent, it will take some time for the system to react before you can continue to spend money, which will also affect ROI to a certain extent.

You don’t need to look at the recommended bid as it is too broad. There is a simple formula: bid = average customer unit price ÷ expected ROI × 1.2. After calculating the bid, create several planned tiered bids. Also, do not bid in integers, but with 2 decimal places.

2. Lycra Orientation

To become a Douyin influencer, first choose to follow, interact with live broadcasts, and interact with products as shown in the picture. Rewards will generally be removed for live broadcast interactions. Then you can select categories and designate influencers. To designate influencers, just enter their names. You can select up to 30.

There is a principle for choosing behavioral interests and Douyin influencers: narrow first and then broad. First, run out orders in a narrow and targeted manner and then slowly add some surrounding categories and related words. In one plan, you can only choose behavioral interests or Douyin influencers, or you can choose both. Neither of them will necessarily have a good effect. The specific effect shall be based on the data that is run out.

After making your selection, check how many people are covered. Generally, it should not be less than 20 million people. The fewer the people, the higher the bid will be required. If the number is too small, it will also result in failure to sell.

Don’t be superstitious or worry about the issue of crowd packages. Toutiao’s current algorithm is already very intelligent. Even without using crowd packages, as long as the method is right, you can still produce good data.

It is best not to enable intelligent volume expansion unless you really cannot escape. If you listen to me, don't use it and just pretend that this function doesn't exist.

For creative classification, you can choose the category corresponding to the product, or you can choose surrounding categories and test which category works best under different plans.

Creative label, all 20 words can be filled in, you can write product words, brand words, functional words, competitor words, target user words, etc.

For example, if I am selling girls' clothing, I can write: children's clothing, girls' clothing, little skirts, little princesses, summer baby short-sleeved shirts, children's short-sleeved shirts, internet celebrity girls' clothing, girls' outfits, high-quality children's clothing, new children's clothing, spring children's clothing, Korean children's clothing, short-sleeved shirts, T-shirts, baby matching, little girls, kindergarten, girls' pleated skirts, cute babies.

Here is a little trick for tags. The first is to go to Taobao and search for your category. There will be many keywords in the title. Just copy them down. The second is to go to Douyin and search for your category and copy the keywords directly.

Finally, you can name the plan and ad group according to your own habits.

Here are the three most important choices for targeting in the entire plan: behavioral interests, TikTok influencers, and creative categories. Carry out different AB tests separately. The core focus of the plan is here. Build different plans for different arrangements and combinations. In the early stage, the targeting should be narrow first and then wide. After it stabilizes, the targeting can be gradually expanded. Finally, when building the plan, you can add a few general investments to quickly get some volume, and the ROI can also be kept in the profitable range.

Don't apply the training slogans of Laman Tongtou to actual operations right from the beginning. You will have to waste a lot of money to find the precise people in your category among the massive crowd of Tongtou. The money is earned by many hard-working people who stay up late at night. If you play like this, people with a weak foundation will lose everything. Be careful!

If you invest in videos, the video itself is more important than targeting. The three factors mentioned above are the most important in targeting. The orientation is narrow at first and then wide, and the number of people is few at first and then many.

Planned Ratio

If the budget is only a few hundred, there are about 3-5 plans to run.

If the budget is only a few thousand, there are about 8-12 plans to run.

If the budget for a show is 10,000 or 20,000, the number of plans running should not exceed 15 or so.

The budget corresponding numbers mentioned above are not necessarily the same, and can be adjusted appropriately according to actual conditions. Because when the budget is small and the plans are numerous, the consumption is dispersed, while when the plans are numerous, the ineffective consumption increases. When creating a new plan, you can turn off a plan with bad data. The standards for turning on and off plans will be discussed in the next article.

2. Planned allocation

As mentioned before, in one plan, you can select only behavioral interests or TikTok influencers, or you can select both, and then select different creative tags respectively. In this way, we can invest in many plans for the same video material.

The initial planned allocation of the account is 28:20, 20% for order placement plan and 80% for testing plan. The mid-term ratio is 55:50, 50% for order placement plan and 50% for test plan. In the later stage, the ratio is 82: 80%, 80% is for order placement and 20% is for testing. The order plan here refers to the same targeted material plan.

In the early and middle stages, run a targeted plan with the highest number of orders and ROI, and then keep copying this targeted plan without changing anything. Let more consumption be used to build data models in this direction, and the model will run more and more stable.

3. Advertising data

After the ad is created, watch it on Qiance

Put the most important data in front:

The total transaction amount, the total number of orders, and the average order amount in the live broadcast room in the overall data

Advertising data includes consumption, advertising ROI, number of advertising orders, advertising transaction amount, and average advertising order amount.

If it is the average order amount of a single product in the live broadcast room, you don’t need to look at this data. Let’s talk about the average order amount of the advertisement, that is, the average order value. Why is it so important? First, it determines your bidding ability. The higher the average order value, the higher the bid. Second, it is a reflection of Qianchuan’s delivery ability and audience accuracy.

Generally speaking, the average order value of advertising is about 80%-100% of the average order value of the live broadcast room. The advertising orders in a multi-product live broadcast room are only part of the orders in the live broadcast room. It depends on the average order value of the people who make transactions in the live broadcast room. The closer the average order value of advertising is, the more accurate the people with average order value in the live broadcast room will be, and the more stable the delivery capacity will be.