Article Description: to create a real case of the explosion of money. |
At the end of November 2010, I received a phone call from Mister brother-in-law of a portal website, let me help him to put more than 10,000 long sleeve T-shirt inventory in the processing, "big winter Let me help you sell T-shirt, this is not stupid B", but dare not offend him, can only bite the bullet to tell But this guy asked me if I could get a shot, and I almost fainted, "this kid must have been kissed by a fool." But go to Taobao to browse a T-shirt data and his shop T-shirt data, found that these T-shirts really hit the possibility of a burst, the other party put forward the condition is that the long sleeve T-shirt into the price is 18 yuan, as long as 18 yuan can deal with it, so decided to draw up a plan to try.
The plan is set, and it will be executed resolutely.
According to the first step of the plan, we first optimized the baby title. First of all, in a simple way, Taobao home search box Search "long sleeve T-shirt female", into the popular baby page.
Figure 2
Then the highest popularity of the top 10 key words of the keyword decomposition, and then put into the Excel list click Sort, to see which keywords appear more often. As a result, we saw that "year words" (such as 2010 2011) appeared 9 times, and the word "T-shirts" appeared 10 times, "" "The word appeared 5 times," Long sleeve "word appeared 9 times," Dozen bottom Shirt "word appeared 12 times," Women "word appeared 11 times," new "word appeared 9 times," Korean version "Word appeared 10 times," Milk silk " The word appeared 10 times and the word "round neck" appeared 2 times, "Low Collar" word appeared 2 times, "Fashion" word appeared 2 times, "Two" word appeared 2 times, "Winter" word appeared 2 times, "Autumn and winter Attire" word appeared 2 times, "thickening" word appeared 2 times, according to these keywords appear frequency, decided to put the number of high-frequency keywords and our t-shirts to match the keyword to form a title, and then the failure to put into the title of the keyword (because the baby title has a word limit) and then form another title.
So the two headings we make up are:
1, 2011 New Korean version of women's long sleeve milk silk T-shirt T-shirt neck and winter
2, two pieces of mail 2010 new women's style long sleeve T-shirt low-age bowling sweater thickened winter clothing
Put these two headings into a through train for testing, by clicking on the number of clicks to determine which title is the preferred title of the buyer. The result of the through-train test was that
Title 1, click 87 Point hit Rate 1.34%.
Title 1, click 105 Point hit Rate 0.41%.
By this data we judged that Title 1 was more popular with buyers, so we decided to use Heading 1 as the title of the baby. Of course, baby. There are other ways to optimize the title, and my "WikiLeaks"-the core secrets of Taobao deals-also has some methods for optimizing the title of the baby.
Next we should optimize the baby's picture, through Figure 2, we find out the rules of the baby picture, so what is the rule of these babies? The answer is as follows;
1, single model 7 70%
2, Product color 4 40%
3, middle position 3 30% (model placement)
4, left and right position 5 50% (model placement)
5, with background Figure 6 60%
6, black T-shirt 5 50% (Model T-shirt color)
Judging from the above data, we optimize the result is;
1, single model, wearing black T-shirt, standing on the left side, the right side show T-shirt color, with background color.
2, another picture adds the word "parcel post".
Through the through-train test, and then see which baby pictures of more clicks, the results of the test with the word "package" pictures of the picture 0.93% clicks, the other one is 0.36% clicks. So we decided to use the message with the words of the baby picture to create a burst.
The third step is a core issue is the price problem, but also the old way to "long-sleeved T-shirt female" keyword into the search box Taobao, then the top 10 of the page of the baby price export, in Excel, calculate the average price of these 10 baby, when we calculate the average price to use the median calculation method. Take a look at the picture below
The table in Excel shows the median price of 41.3 41.99, respectively. After the average price is 41.6, then 41.6 of the price is not what the buyers like? We're not sure what to do. Smart readers will immediately remind me to test, congratulations, correct. We have tested the highest, middle and lowest three kinds of prices (88 41.6 19.99) in the through train, noting that the test results we want are not clicks, but conversions, so what are the indicators that we should assess the conversion rate? What is the number of deals? At this time with the number of transactions to assess indicators contingency too large, with Wang Wang's question volume as a test indicator should be more accurate. The results of the three prices came out,
1, Price 88 point hit 66 question number 7 question conversion rate 10.6%
2, Price 41.6 Click 93 Question Amount 16 question conversion rate 17.2% deal a sum
3, price 19.99 point hit 97 question number 3 question conversion rate 3%
3 of the above data do not need me to explain it, the best performance is 17.2%,41.6 is the price we want to use.
The fourth step is to describe the part of the baby, this is about your baby is able to deal with the key, but most of the small and medium sellers did not do this part, do not even know how to do. In the optimization of the baby description, first of all the above 10 popular baby's web page open, to see how their baby is described. According to the description of the 10 baby, we found their description of the law, the 10 of the description of the popularity of the baby there are about two rules, according to these two laws, we established a model of two baby description. The pictures are as follows
According to the models described by these two babies, we made two baby descriptions, or put them into a through train for testing, and see which baby describes the conversion rate higher? So what does the baby describe as a reference to conversion rates? We put the average stay time of the baby page and Wang Wang's question as a conversion rate of assessment indicators, so baby description of the page stay time based on what? The solution is simple, we've got three employees to read, baby. Description page, the average reading time of three people is 3 minutes 40 seconds, that is, an average of 180 seconds to read the baby description of the page, so we will be in the 180 seconds of time as a transformation to evaluate the baby description page. The results of the test are as follows
1, click 94 Times 180 seconds above the stay time 27 people Wang Mong question 19 people
2, click 87 Times 180 seconds above the stay time 34 people Wang Mong question 22 people
The second figure is clearly more convincing to buyers. Then the second baby description is perfect? Obviously not, in order to make the description page layout more reasonable, more in line with the buyer's browsing habits, we separate the Baby description page 2 of the content of the copy, made a separate Web page, to carry out multivariable testing. Finally, we found that when the model picture and details + Picture position interchange, quality and detail picture placed on the left side of the text on the right side, multivariable test results are the best. The pictures are as follows
So our baby page is also optimized (hint: If you do not have the conditions to make the site, it is recommended that you do not carry out a multivariate test of the Web page).
Through the 4 steps above, the results of my optimization are as follows,
Baby title: 2011 New Korean women's long sleeve milk silk T-shirt t-shirt neck/winter Bottom
Baby Pictures:
Baby Price: 41.6
Baby Description:
The above four steps of the test and modification time we used about 10 days, the optimization of natural flow after the click Volume increased by 30%. So we can go to the real delivery through the train to build our money. So what's the effect of the through train? Can this t-shirt be made into a bomb? "Here is the time to witness miracles!" Please continue to pay attention to: "Build the explosion" The Real Case (ii)
Tips:
1, this case is 2010 case, because the time is too long, some picture data is not complete, so some pictures use now picture to replace.
2, the case products in November 2010-April 2011 has been sold out, so please do not take a seat.
Author: Cyclone 1968