A 10,000-word article sharing the complete practical details of e-commerce testing, my personal practice summary version (Part 1)

In the world of e-commerce, product selection and market acceptance testing are the keys to success. This article will take you deep into the complete practical details of e-commerce product testing, from theory to practice, revealing how to verify whether a product meets market demand through meticulous process management and advertising.

Friends who have never done e-commerce, or those who have little experience in e-commerce, may have heard that e-commerce requires testing of models, products, and images. However, most people are actually ambiguous about how to do this process from beginning to end.

There are some articles about model testing on the Internet, but I read a lot of them and felt that the descriptions were rather vague, and I still didn’t understand them after reading them.

I also listened to a lot of e-commerce courses (yes, I also need to keep learning). Some e-commerce courses talk about product testing quite well, but the methods of these courses are mainly aimed at people with rich e-commerce experience and proficient operations. When ordinary people listen to them, they are basically confused.

For example, I listened to a course taught by a big e-commerce tycoon. His way of testing products was to build an advertising plan, put 100 products in it, and then add a few hundred keywords, and whichever product came out first would be the one that was tested.

They are quite skilled in using this method because they have many stores, many categories, and many products, which requires large-scale testing operations.

This method is very suitable for e-commerce veterans and mature teams. But if ordinary people really implement it this way, they may die suddenly at the beginning, because just finding 100 products is already tiring, and they also have to dig out hundreds of keywords.

I think ordinary people can't carry it out.

Some e-commerce experts talk about testing, promotion, and label weight at the same time, mixing multiple dimensions together. The content is correct, but many people don’t even understand the backend operation of promotion. At this time, if you talk to them about labels, weights and other very conscious things, their brains will be completely entangled and they will not be able to clearly understand what the core focus is. So after combining many courses and doing a lot of practice in my own store, I achieved some good results.

I feel that I have come up with a relatively concise, clear and easy-to-use simplified version of the testing process.

Please note that this is the simplified version.

I am sharing this with you today, hoping that everyone can have a comprehensive and clear understanding of the e-commerce testing process.

I would like to point out that I work in the clothing category, so I prefer to say "testing models". Most e-commerce people also say "testing models", and a small number of people say "testing products". But in fact, the two words are the same thing, both of which involve testing products. For the sake of convenience, I will use the word "testing models" below.

1. The importance of testing

First of all, one thing is clear: testing the model is not necessarily important, but the nature of the reaction to the testing is very important.

What essence does the test reflect?

It means whether you can clearly grasp, observe, and verify whether the market will accept your product.

And this link must be pre-planned, the clearer in advance the better. Only when you are sure that the market is willing to accept your product, you can promote it and prepare stocks according to demand.

Why is it difficult to run an offline store? Especially offline clothing stores, restaurants, and small commodity stores are difficult to run because they cannot grasp in advance whether the market in this area will accept their products.

There was no offline testing process, no advertising tools, and no way to test in advance. Only after the store was opened and the goods were purchased could feedback be obtained. But by then, a large amount of money had already been spent on rent, renovation costs, deposits, and purchase costs. As a result, it was discovered that the market did not accept its products, the goods could not be sold, and no customers came to the store. After a few months, it had no money to pay the rent and went bankrupt.

Many people who create original products also make this mistake.

Many Internet elites, white-collar elites, and workplace elites started e-commerce businesses with their savings. They looked down on the products on the market and decided to create their own. They did a lot of research, which they thought was scientific, and added their own aesthetic tastes, understanding of the products, and understanding of the market. These understandings were not tested in the market, so they directly started to stock up, release, and advertise in large quantities. In the end, they found that they could not sell the products, and all the goods were stuck in their hands. They lost their savings and had to go back to work.

Therefore, testing is just a means, and we still need to constantly question ourselves:

• Can I clearly grasp the needs of the market and consumers? Do I have any successful cases?
• Will the market accept my product? Is there any data to support it? Have you conducted a small-scale test?

If you don't have these, you need to test the product first. After you get a lot of data to support your decision, you can take the next step.

2. In some cases, no testing is required

I would like to emphasize that it is not necessary to test products when doing e-commerce. We should not test products just for the sake of testing them.

In some cases, you don't need to test the money.

For example, if you have no money, and your promotion expenses are very little or almost nothing, and you are asked to spend dozens of yuan a day and a few thousand yuan a month, you are reluctant to do so, because you have no money.

I am not mocking you, I am just stating the reality. Most ordinary people who do e-commerce have no money.

Then you don’t need to test the money. What is there to test? You have no budget and no money, so you have to think of other ways.

For example, you go to Xiaohongshu or Douyin to shoot a short video of a product and test it, and then you say you don’t know how to shoot a short video, you can’t do this, you can’t do that, then you should just go to work honestly!

For example, if you don’t even know the basic advertising promotion methods and operations, then you don’t need to test the product.

We test models in order to find good products, and then use promotional tools to promote them. If you can't even use advertising promotion tools, then there is no point in testing models. Even if you find a good model, you won't be able to promote it later!

For example, if you are a great product manager, your aesthetics and product selection ability lead the market and your customers, and you are a great designer with successful cases, then you don’t need to test models, you just need to focus on product development;

Don’t think these people are uncommon. In fact, some of the internet celebrities, influencers, or experienced people who have been working in a certain industry for a long time all have this ability.

They have been doing this for so long and have a deep understanding of the industry that they can probably tell whether a product will sell well just by looking at a few of them.

Therefore, I also suggest that you should research the products more and accumulate in-depth experience in an industry. The more experience you accumulate, the less you need to test the products. Your vision will be more accurate and you will be a human product testing machine.

For example, if the product you sell is already selling well in other stores or competitors, then there is no need to test the product. Your competitors have already proved that it is a good product and is acceptable to the market. We only need to test the other parts.

For example, if a certain product is already selling well at the supplier and the shipment volume is huge, then there is no need to test the product. The supplier is shipping a large amount of the product, which means that there are many distributors selling the product, and it must be a product that everyone thinks is good.

For another example, you still need to compare and judge the product, conduct some competitive comparisons, compare it with products of the same industry in terms of style, price, pictures, etc. Once you have a rough comparison and think they are almost the same, you can test it again. If it is obviously bad, unacceptable, and a very ordinary product, there is not much point in testing it, and you would waste your money.

In short, the purpose of our product testing should be clear, to find products with high market acceptance among a bunch of products, or to understand whether the market is willing to accept the products we make. If these things have been proven, there is no need to test the product again.

3. Are there some methods that can replace on-site testing or serve as a supplement?

Our testing of funds does not mean that we must use in-site advertising to test funds, or that we must spend money on advertising. If you have other capabilities, you can also use them as a substitute for testing funds, or complement them with in-site testing funds. In this way, the advertising costs of testing funds will be greatly reduced.

For example, if your store has been open for a long time and you have a lot of loyal customers, every time you launch a new product, you will get a group of loyal customers to come back and buy it. A certain product performs particularly well among old customers, and seven or eight, a dozen, or even dozens of pieces are sold as soon as it is put on the shelves. Then you don't need to test the model, you can just push it directly, it must be a good model, and old customers themselves represent market demand.

For example, if you are doing live streaming, you don’t need to start a special advertising plan to test the models. You just need to promote new models every time you broadcast live and monitor the data. Testing the models has been integrated into your live streaming, so there is no need to test it deliberately.

For example, if you are a blogger or KOL on Xiaohongshu, you have a good self-media account, and you have certain data when you post product videos. Then you can directly test the product data through short videos or notes on self-media, which can also replace the test model.

Why do online celebrity stores have such a great advantage? When it comes to testing products, many online celebrities have Weibo and Xiaohongshu, so they can preheat new products and test products in advance, quickly determine the main focus, and reduce time costs.

Therefore, don’t be too dogmatic about product testing. As long as you can use data to prove that your product is a good one, you don’t have to be limited to in-site advertising.

4. Product preparation before testing

Testing a product actually means advertising the product and obtaining feedback data.

In order to get better feedback, we need to consciously adapt advertising during the initial planning and product launch, so that we can get better results and reduce promotion costs.

The first is that the products of the general test advertising plan should preferably be of the same category.

For example, I put up 20 products this time, including 6 sweaters, 5 pants, 3 dresses, and 6 knitted tops.

When I test styles, I will try my best to include sweaters and base shirts in an advertising plan for testing.

Because there will be keywords in the advertising plan, for the same category, the keyword concentration is high, there are many adaptations, it is easy to increase the weight, the effect will be fast, and the efficiency will be improved, which is equivalent to saving promotion funds in disguise.

This is why the women's clothing category emphasizes the importance of category planning. In the initial stage, categories cannot be scattered, but should be concentrated on 2 to 3 categories. This is also to create efficiency concentration for subsequent promotions.

Secondly, when I test the models, I try to ensure the consistency of the pictures.

Because e-commerce is all about selling pictures, some products are very good, but if the pictures are not good-looking, the good product may be buried.

We try our best to ensure that the style of the images is consistent. This also has the advantage that if we find that the click-through rate of the images is not good during testing, we can perform unified optimization in the next round.

The most worrying thing is that some pictures are good but the products are bad, while some products are good but the pictures are bad. The data obtained will be very confusing, making it impossible for us to analyze it.

Then, I still recommend that when you first start doing e-commerce, you’d better focus on the mainstream price range and the low price range.

It is not necessary to set low prices, but at least the products should be within the mainstream price range.

You sell lunch boxes, and lunch boxes on the market are priced between ten yuan and dozens of yuan. Don't go against the grain and make a lunch box that costs hundreds of yuan. Consumers cannot afford it. Many people make very good products, but they are also very expensive. You cannot get accurate data when testing the product. The product can attract a lot of traffic, but there is no conversion. At this point, you cannot judge whether consumers are not interested in this product because of the price or the product itself.

So you can see that in order to create good conditions for the testing process, we try our best to use the control variable method to keep some data as uniform as possible and keep some data within the standard range of the market, so as to avoid abnormal influences.

5. Specific Testing Tools

Taking Taobao stores as an example, what tools do we use to test products?

We use the through train to test the products, which has now been renamed the [Keyword Promotion] tool to test the products.

Because the purpose of testing a product is to use the strong demand of customers to prove the product, then this strong demand must be based on the user's search behavior.

For example, if I sell water cups and want to test water cup products, then when users search for the word [water cup], I hope that the advertisement can show my water cup products to users to see whether users will take a series of actions towards my products. Therefore, the [Keyword Promotion] module is an active behavioral advertisement based on user searches, which truly represents the strong demand of customers for this product.

6. How to create a testing plan

Since many people have not done this, I will quickly and easily teach you how to build a test plan. Friends who have not done it can just understand the process. If you don’t want to read this part, you can skip it.

For those who have not operated it before, it is important to experience the feeling and the actual operation experience from the front line.

Step 1: Product Selection

Select [Keyword Promotion] - [Custom Promotion] - [Custom Product Selection]

When testing a product, we must test the product we want to test, so we add the product we want to test.

The number of products will definitely not be large. I believe that none of us have the ability to produce dozens of products at once, so we can initially limit the number of products to be tested to less than 10.

Step 2: Budget Setting

We choose [Manual Bidding] and set the budget to [30 yuan per day]

After the ad is launched, we will bid for the keywords we want to invest in. In order to control costs, we must do this bidding manually.

For those who are just starting out, spending only 30 to 50 yuan a day on testing is enough.

So as long as you control the amount of testing, it will cost some money, but not too much.

Step 3: Select keywords or word packages

After adding products, the system will automatically give some words based on our products. I added some sweatshirt styles, so the system gave me words about sweatshirts.

We can check whether these words match the products, select them if they match, and delete them if they don’t.

If you make a product, for example, you make a certain product, such as a foot bath bag, I don’t know why I would say this product, and you find that the system gives too few words, or they do not match your product, then you should turn off all the words, and then select the traffic smart word package given by the system, and use the words recommended by the system to test.

Of course, the above word selection can be said to be the simplest and most direct method, because they are all recommended by the system.

Experienced e-commerce operators will accumulate a large number of keywords, and will accumulate a large number of precise keywords for the categories they sell. When they test their products, they will choose from the keywords they have accumulated. But I think this is unrealistic for many novices. You can gradually accumulate your own vocabulary in the subsequent promotion process.

In the early stage, it is enough to use the system's recommended words. After everyone gets the hang of it, you can gradually go deeper.

Step 5: Select the audience

In addition to selecting keywords, advertisements are also displayed to different groups of people. This is because there are many groups of people on the e-commerce platform. It does not mean that if you bid for a certain keyword, your advertisement will be displayed to everyone. Your advertisement will only be displayed to a small part of the selected group of people under your keyword.

So what group of people do we choose?

For new stores, just select [Smart New Customers]

For old stores, select [Smart New Customer Attraction] + [Long-term Store Value Group] + [Visitors Who Have Purchased Store Products]

In short, it is the [intelligent new customer acquisition] crowd. The system helps us recommend products to more new groups of people. The store crowd is to show your products to old customers. In fact, at this point, a small closed loop of the testing plan has been completed. Add the product, add the keywords, and add the crowd. Your product will be displayed to a part of the people who have demand for this keyword.

These people will come to see your product, whether they are interested in it, and whether they have any interactive behavior. Through these behaviors, you can judge whether the market recognizes your product.

If you are rich, you can add more groups.

Of course, this plan has not been completed yet. There are still a lot of details that need to be adjusted. After the test plan is built, how will they perform, how to optimize it, and how to follow up on the plan. We will talk about it in the next article, so stay tuned!

(To be continued)