Use "subdivision" to find out the latent factors that affect the baby's flow

Intermediary transaction http://www.aliyun.com/zixun/aggregation/6858.html ">seo diagnose Taobao guest cloud host technology Hall

A lot of friends asked me to write about how to use data for marketing analysis, since you like to see the case of Taobao analysis, just to catch up with a recent real case, an old customer to me, their C store traffic has been very stable, on March 6, the traffic suddenly dropped, they are anxious to find out why. This is a more typical case, but also very interesting, I hope you read carefully, then the first photo of it.

This is a very common quantum graph, all the flow is very smooth, also relatively natural, March 6 suddenly fall (with red circle Mark), March 12 suddenly rise. Today we analyze the reasons for the March 6 decline, March 6 good This day's flow of the decline, what happened this day? Can we find out the reason for the decline? In order to find out these reasons, what are the numbers we want to subdivide? So how to find out the reason for the decline in traffic by subdividing? This is also today's large Taobao small sellers have been puzzled by the reason, and even some big sellers for these problems in confusion. There is a very simple way to find these reasons, I can compare March 6 with the day before, and then compare it with the one, I can make a preliminary judgment. Look at the bottom three pictures.

March 5 Visitor Source data

March 6 Visitor Source data

March 7 Visitor Source data

From these three, can you see where the data is changing? If we can't make a judgment right away, then we're going to use Excel to do a simple perspective model (the next picture).

Excel Perspective Model

Through this simple model we can see very intuitively, March 6 this day is because the class purpose traffic is reduced, causes the shop overall flow to drop. This day is in that period of time the flow dropped where? We can use the same method, and then to the flow time period analysis, to see the following time period flow chart. #p # subtitle #e#

March 5 time period guest map

March 6 time period guest map

March 7 time period guest Map

These three diagrams may still be the confused you see, but when we follow the above method and do a simple perspective model of Excel, we can see clearly which time period the traffic is falling (the following figure).

Time period Excel Perspective model

Through this Excel model, we can clearly sentence the segment, the visitor's traffic is at 12 o'clock noon starts the noticeable descent, but continues to the night 24 o'clock. So March 6 is the day of that baby or what the baby's traffic is declining? Look at the following 3 Treasure flow chart. #p # subtitle #e#

March 5 Baby Visitor Map

March 6 Baby Visitor Map

March 7 Baby Visitor Map

If we use the same method to make an Excel perspective model of these three graphs, we can see more clearly (below).

Baby Excel Perspective Model

Through this Excel perspective model. We can make it clear that March 6 is the day because baby A's flow down, affecting the overall flow of the shop. So baby a why did the traffic suddenly drop after 13 o'clock on March 6? What the hell is going on this day? Through these characteristics, we made two judgments.

Think of one, when clicking on the category page, sometimes appear is a new baby page, sometimes a popular page (figure).

If you go into the category today, show the popularity of the page, and your baby also appear on the popular page, then this baby today's clicks on the high, as shown today is a new baby page, and your baby did not appear on this page, then the number of clicks today is small. And we this baby just display in the category of the homepage of the popular page, the answer is March 5 this day into the category, showing the popularity of the page, and March 6 this day into the category page, showing the new baby page. And baby A doesn't appear on this page, so the traffic is decreasing this day. and March 7 into the category page, the display is a popular page, so that our click volume again.

Think two, appear this may be March 6 the day baby A after 13 points were artificially off the shelf, or by Taobao punishment, after we check with customers, both of these situations are denied.

So the answer comes out, and the answer is the result of our analysis of "thinking one." is the March 5 and March 7 These two days, into the category after the display is a popular page, and baby A also appears on the homepage of the popular page. The March 6 category page shows the new baby page, and baby A does not appear on this page (only the baby in this world will appear on this page), so in this genius led to a significant drop in traffic.

Probably most of the friends think that the analysis to the end of this, if you want to think this is a big mistake, such errors are often the site operations analyst, can analyze problems, do not solve problems, resulting in frequent contradictions between the boss and analysts. So how do you solve the problem in this case? I believe a lot of smart friends will give all kinds of answers, but the way I use it is that when this happens, we can directly promote the category of the through train to the first page, the eighth place is OK. Of course, if the flow of traffic has no impact on your trade, you do not have to pay attention to, if the impact on your transaction, you must let the through-train category to promote access to the home page, of course, the prerequisite is that you must spend a little yuan. Saharan

This case surface does not have any wonderful place, but it tells us as long as we through the subdivision can put each layer open, finally found the reason, encountered this kind of problem must analyze all the data, do not go to "guess", because the data is not a riddle.

Please specify from the generation network author _ Cyclone 1968