这篇文章的中文版以前在WAW (网站分析星期三)的微信公众号发布过的。
- 中文版的翻译者是Lucky Chen。
- 英文原版是我写的，以前在Clickz.com上发布过。
中文版和英文版，这次我都一起放出来。
现代网络分析师所扮演的角色
四件事
作为网站分析师，你需要做以下四件事：
- 趋势报告及数据报告
- 分析当前线上营销获取策略以及探索新机会和/或新策略
- 理解网站用户的行为和经历
- 持续关注趋势和细节
三个阶段
我们可以将现代网站分析师的工作分为三个阶段:
- 数据收集
- 数据提取
- 数据处理及展示
第一阶段：数据收集
假设在线业务大多数都是在你公司官网上进行，那么大部分线上营销、搜索营销和用户行为活动的相关数据，都可以通过：
网页日志文件收集 — 你必须精通网页日志文件收集数据的原理，并且知道哪些数据是可收集的。网站日志文件可以"记录"所有用户在网站上加载的文件，因此你可以轻易地发现网页的哪些"部分"没有响应用户的请求。网站日志分析参考：http://cn.analyticsbook.org/web-server-log-analytics/
网站分析 — 全球大部分网站都使用分析工具。网站分析工具一般具有图形界面，可以快速显示用户的数据趋势。所有数据可以以表格、文本文件甚至是 PDF 文件的形式下载到本地。
利用网站分析工具收集用户数据前，需要安装基础设置来追踪数据。通常要插入一些JavaScript的追踪脚本或者在网站所有HTML 页面插入一些1*1像素的脚本。如果你需要收集的用户数据超出默认设置所收集的用户数据，需在常规追踪脚本外安装高级追踪脚本。
这些包括：
- 国外的免费工具：谷歌分析
- 国内的免费工具：百度统计、CNZZ 统计、腾讯分析
- 国外的付费工具：Adobe Site Catalyst
- 国内的付费工具：99Click
搜索营销平台 — 有些网站利用第三方搜索营销管理和跟踪平台，比如Kenshoo, Marin Software 或Adobe AdLens（以前叫Efficient Frontier）。这种情况下，需要在网站上安装该平台的跟踪脚本，才能收集信息。
注：提到了超过200个网站分析与APP分析工具的大全：
http://cn.analyticsbook.org/the-big-list-of-analytics-tools/
第二阶段：数据提取
数据收集好后则要提取信息。第一阶段收集的原始数据需要转换成报告，主要有两种形式：
-
常规数据报告：这些报告通常有一定的时间规律，按照报告性质可按每天、每周和每月生成。报告根据收取人的要求分成不同等 级，比如执行官可能跟公司主要部门开会时需要展示关键收入的报表。运营经理可能需要中级数据报告去追踪他们组负责的产品 的"潜在问题"。
-
即时数据报告：这些报告并非常规需要。通常即时数据报告是为了回顾所有一次性线上活动的效果。如需深入了解数据所代表的 问题，比如为什么过去两周KPI 数字有所下降，这时也可用到即时数据报告。
商业分析（BI）团队可能已经处理过原始数据并把数据转化成可读报表。一些商业分析数据仓库系统可以制作这类报表，比如Cognos或其他相似的数据平台。通常公司都是用这些来做常规报表。即时报表则需要更高的处理时效。
要想从MYSQL数据库、Oracle数据库或者其他数据库直接提取数据，你需要用到SQL语句。并非收集的所有原始数据都需要进入到数据库或者数据仓库。因此你要花很多时间抉择到底要处理哪些数据。比如：
-
如果你选择了一个免费的网站分析工具，比如Google Analytics，那么网站数据则全部存放在谷歌的服务器上。你可以在GA后台 上选择以表格形式把数据报告下载到本地，或者通过谷歌提供的API 直接提取数据报表。
-
搜索营销平台的数据有可能也存放在供应商的服务器上，此时你只能以表格的形式提取数据。
第三阶段：数据处理及展示
Excel：在制作漂亮数据图表报告之前，你需要对SQL 语句提取的数据做一些处理。通过Excel 做漂亮报表相当耗时，因此，熟练并专业地使用Excel 各种功能和公式能够极大地提高工作效率。例如利用透视表整合数据、利用Vlookup函数合并数据和利用“月”、“日”等公式处理日期等。
我写的英文原版：
The Role of the Modern Day Web Analyst
As a web analyst, your role requires you to perform four major tasks:
- Trend and data reporting
- Analyzing current online marketing acquisition strategies and exploring new opportunities and/or new strategies
- Understanding on-site visitor behavior and experiences
- Staying connected with the trends and the details
Let's go through the three phases for the modern day web analyst:
- Data capture
- Data extraction
- Data manipulation and presentation
Phase 1: Data Capture
Assuming your online business mostly happens on your company's websites, most of the online marketing, search marketing and user behavior activities can be captured with:
Web Log Files – You must be very familiar with the principles of how web log files capture data and what data is available. Web log files give you the ability to "record" all the files that were loaded by the user when they accessed your websites, and you can easily see which "components" of your websites aren't responding to user requests.
Web Analytics – Most websites globally use some analytics tools. A typical web analytics tool provides graphical user interfaces (GUI) and allows you to quickly see the data trends of your users. Reports can be downloaded as spreadsheets, text files or sometimes even as PDF files.
Before a web analytics tool can capture users' data, you are responsible for implementing the required tracking setup. This normally requires inserting some JavaScript tracking scripts or some 1×1 pixel scripts onto all HTML pages of your websites. If the objectives include capturing more than the default amount of user data, then you are required to implement some advanced tracking scripts on top of the regular tracking scripts.
These include:
- Free & global: Google Analytics
- Free & local (in China): Baidu Tongji, CNZZ Tongji, Tencent Analytics
- Paid & global: Adobe Site Catalyst
- Paid & local (in China): 99Click
Search Marketing Platforms – Some websites make use of third-party search marketing management and tracking platforms such as Kenshoo, Marin Software, or Adobe AdLens (formerly called Efficient Frontier). You are required to implement the platform's tracking scripts onto your websites in order for the data capture to work.
Phase 2: Data Extraction
Once the data is collected, the next phase is to extract the data for the end users. Raw data collected in phase #1 should be converted into reports that are for two major purposes:
-
Regular Data Reports: These reports need to be received on a regular basis, whether that's once per day, per week, per month, depending on the report. These reports are categorized into different levels depending on the receiver, so an executive would need high-level reports showing key revenue numbers for each major division of the company. Operational managers would be looking at mid-level data reports that allow them to track "potential problems" in the products that team is responsible for.
-
Ad Hoc Data Reports – These reports won't be processed regularly with any fixed intervals. Normally, ad hoc reports are required for review purposes for any once-off online campaigns. Ad hoc reports are also required when you need to dive deep into the data in order to figure out problems such as why certain KPI numbers have decreased over the past two weeks.
The business intelligence (BI) team may already have processed the raw data and have it converted into reports which are readable. The reports can be obtained under some BI data warehousing systems, for example Cognos or other similar data cubes. These reports can form a large part of the regular reports to a company. Ad hoc reports often require quicker turnaround time.
It is time for you to utilize your SQL query ability in order to extract data directly from the databases whether they are MYSQL databases, Oracle databases, or other databases. Not all the raw data you have captured goes straight into your databases or data warehouses. For this reason, a large of amount of time goes into then working on post data manipulation. For example:
-
If your choice of web analytics is a free tool, such as Google Analytics, then your web data is all sitting on Google's servers. Your options are to either download the data reports onto spreadsheet formats through the Google Analytics online interface, or extract the data reports through Google's APIs.
-
The data from the search marketing platforms may be sitting on your vendor's servers, and you can only extract the data into some spreadsheet formats.
Phase 3: Data Manipulation & Presentation
Excel – Before you can present pretty graphical reports, the data extracted through ad hoc SQL queries will need to be manipulated. Creating pretty reports through Excel can take a lot of time, therefore, having expert knowledge on how to get the most out of Excel features and formulas will only improve efficiency. These could include: aggregating data with Pivot tables, merging data with Vlookup, manipulating dates with functions like "day", "month", etc.
相关阅读
假如对数据分析技术有兴趣，推荐阅读我写的书，可以免费网上阅读数据分析技术白皮书：
我的个人微信号二维码
需要联系，就请扫二维码。
网友评论