E-commerce is developing faster and faster, and the trend of this industry is changing more and more quickly

A few days ago to see a data enthusiast, we have a consistent view: e-commerce development faster and faster, the trend of the industry has changed more and more quickly; for E-commerce company bosses, want to walk in front of the trend, must learn data driven.

Fortunately, this year, e-commerce operators on data analysis has been paid attention to, even by the pop taobao sellers began to recruit data analysts, not to mention some larger e-commerce companies.

But it also worries me: now is not the lack of data, but too much data!

According to statistics, in today's Internet, every 60 seconds will produce 100,000 microblogging information, 4 million search, Facebook 500,000 times contact. I believe that some of the larger e-commerce companies today will collect some behavioral data, but what does the behavioral data have to do with business data (such as trading volumes)? The vast majority of companies today do not make good use of these thousands of scattered data.

Need data logic, more business sensitive

Tell a funny story first. One day, LinkedIn suddenly found that Lehman Brothers had more visitors, but did not delve into the cause, and soon after, Lehman Brothers announced their collapse. In the first one months of Google's announcement, I found some of the most rarely seen Google product managers online in LinkedIn.

Just think, if LinkedIn is focused on the analysis of a listed company executives looking for work data, is it very commercial value? I believe that there are a lot of websites that do not know how to collect these deep data (the bottom datum, editor note), only stare at some simple surface data, even hold the shark fin as radish.

This story, just to tell you that the data in the Internet, need to use a commercial perspective to analyze and correlate, only value.

Today's E-commerce company's data analyst, must have from the boring data to see the ability to unlock the market password.

For example, when a business-minded data analyst discovers that the demand for baby cars on the site increases, he can basically predict which associated products will be sold.

For example, similar to traditional stores, the site's products play a different role. Some products are to make money, some products are to promote, and some products are to lead the flow, different products in the site placed on the location of course is not the same. This can also be found in the data.

A commercially sensitive data analyst needs to know what data to use to drive a company to achieve its goals.

Another example, two new Consumer-to-consumer platform competition, the focus is not simply trading volume, but how many new seller each day (sellers, editor note) came in and sold how many things. Because the core competition at this stage is popularity, not real trading. If the new seller comes in and doesn't sell, it's just that the old seller are growing, and even if the final volume is growing every day, there is a problem.

For example, a company that has just entered the market and has already captured most of the market, has a different corporate goal. The former value flow to earn popularity as the goal, but the latter significance of traffic is not so big, mature company focus is to see the transaction, conversion rate and turn heads.

The current data analyst is mostly statistical, and a bunch of data is there, and everyone is good at how to calculate regression and how to draw functions. But the math talent is less business conscious, does not know what this data means to the business, see a lot of data who and who has a relationship, also do not know what to use logical analysis, not to act as the boss's eyes.

A few days ago met a boss, he said his staff to show him dozens of pieces of data every day. I ask, is not the more data the more trouble. He said I had a quick spot on him, because the data analyst who came to him only handed the data to him, but didn't tell him about the relationship between the data and the business.

You say, a company CEO, see dozens of data every day, what PV, PU, UV and so on, they have the energy to interpret it? For them, just know: Is there a problem with the company? What's the problem? Any new discoveries? What do you need to do? That's enough.

I understand these issues as the "dashboard" in the data world, such as how web traffic comes in the pop-up rate can be presented in the dashboard. You drive, if the water temperature is too high, the instrument panel lights the light hint. Similarly, in E-commerce transactions, some data can also be used to form a "dashboard."

Therefore, the data analyst should not be a simple math problem.

Behavioral data and business data, pushing each other

A good dashboard, there are good conditions and bad situation, the dashboard will be prompted. What constitutes a "dashboard" is the logical relationship between behavioral data and business data.

I'm getting used to this kind of appellation: front-end behavior data and back-end business data. The preceding data refers to the data of the response user behavior, such as the traffic volume, the browsing volume, the click Stream and the site search, and the latter data is more focused on commercial data, such as transaction volume, ROI (conversion rate), LTV (Life time value).

Some people are concerned about behavioral data and some care about business data, but few people link behavioral data to business data. People tend to simply look at one end of the data. The domestic small famous website CEO, also sees only one result data every day: The website today's volume is how many, sells the number of pieces of product.

But people who look at the data are going to understand that every data, like a star in the dark, is a network of contacts, and that simply clicking one of the data will drive some other data to change.

Everyone is more concerned about the site user base, as an example of this.

One day, a website found that their front-end registration increased a lot, the number of visits also went up, trading volume did not go up, Buthen. In fact, this is also a common problem of many websites, every day many bosses are thinking about this issue.

What's the reason?

At this stage, people who are on the front end of the internet only know the number of clicks and so on, and rarely ask the back-end business data, such as who has been repeatedly buying? Who affected 5%-15% of the core users to come in and buy things? Who is doing positive and negative communication to the website?

And the operation of the site back-end trading link people only know to sell things, and rarely asked the front-end data. If a customer in the site to stay on average 15 minutes or 30 minutes, this is the relationship between the future repeat purchase? Does a customer enter the site community and do not enter the community, the impact on the production volume?

This belongs to the core user group, the reason, to a large extent, because the behavior of data and business data docking to see.

So, the data is fragmented, and no one knows the relationship. As the site's decision-makers, do not know the site's core user group behavior characteristics, also do not know how to stimulate the increase of the core users, not to know from a user to enter the site to go out, which link is need to dredge.

Of course it's just a glimpse.

A platform operator, the response user behavior of the front-end data and back-end business data tens of thousands, sellers and buyers are tens of thousands, which data on the front end of the entire Web site on the transaction volume has the greatest impact, as long as the front-end data on the drug, will stimulate the back-end data increase; The back end which transaction data is relatively high, is clear from which channel, the main contribution user is who, the website product design must incline to them, to them better, so will have the channel front end "The conversion rate" and so on key data promotion.

If the core user base of a website grows at a rate of 10% per month, it is also a strange thing.

Unfortunately, today many e-commerce companies, every day in the "happen" game: Today recommended a products, tomorrow to remove a home products; today do a low price promotion, tomorrow and do offline activities. Changes in these decisions, with no dashboard instructions or good data monitoring, are clouded by coincidence.