The eternal questions for online retailers have a simple answer these days. How can you identify an emerging shopping pattern? With your website analytics tool! How can you reduce your costs and manage your website better? With your website analytics tool! What’s the best way to build meaningful relationships with your customers? Well, you know the answer…
E-shopping analytics is necessary because without data, you can’t make the right decisions for your business. Let’s check if you’re ready to analyze your e-commerce data. And If you already are, let’s see if you’re doing it right.
What you have to know to get started with e-commerce data analysis
First, let’s take a helicopter view and look at the whole data analysis process.
For a big online retailer and a little bakery selling their buns online, data analytics means the same thing — creating sensible metrics that reflect the status of the business and getting insights on consumer behavior.
Thanks to analytics, you can find bottlenecks in your sales and marketing activities, change your strategy faster than your competitors, and spend your time and money on the most efficient activities to take the market by storm.
Stages of data analysis
To make your data analysis meaningful, you need to:
1. Understand what data sources and what kinds of data are available to you
Get a whiteboard or open a new document and write down all your sources of customer behavior and advertising data: online chats, support forms, etc. Also, check out the formats this data comes in, paying attention to those services that don’t have automated data export functions and don’t have APIs for integrating with your analytics tool.
These services require more of your time and manual work, which isn’t the best news.
At this stage, you have to define all the metrics that reflect the most essential business events and situations your business may find itself in as well as those metrics you want to measure in your analytics system.
2. Collect your data
This is the stage where a good technical setup will benefit your business. If you don’t know anything about tracking tags, cookies, browsers, etc., find someone who does and make everything neat and clean. Try to collect all available data. You’ll use all of it once you have enough experience.
3. Process your data
Data processing means merging separate data streams into a single dataset and preparing this dataset for the next stage — data cleaning. During the data processing stage, the reliability of your data processing software means a lot.
Schema for collecting and processing data for Fabelio, a multi-channel furniture store.
Source: OWOX blog
4. Clean your data
Your data will be really “dirty” at the beginning — some data will get lost, some will be erased, some format conversions will go wrong, you’ll learn to make backups, etc. You’ll need to correct errors, delete duplicates, and heal your dataset to make it ready for data analysis.
If you skip this stage, all your previous and further work won’t benefit your business. You’ll just waste your time.
5. Data analysis… at last!
At this stage, all of your efforts will be repaid. But don’t get us wrong: it’s still a lot of work. You have to rely on your metrics system and see how it works with real revenue and cost data. This is the stage for reporting, visualization, shifting results, and finding insights.
Types of e-commerce data analysis
Depending on the sophistication of your data analytics, you can conduct different kinds of analysis with your e-commerce data:
- Descriptive analysis for revealing what's happened
This is the basic level of data analysis. It gives you an understanding of what happened in the past based on tracked events and additional data. Descriptive analysis can provide you with:
— A KPI dashboard
— Weekly and monthly revenue reports
— Overview of sales and lead generation
- Diagnostic analysis for understanding past events
This type of analysis shows you why things happen by revealing hidden trends and patterns in data and how they correlate with customer behavior. With diagnostic analysis, you can:
— Find out why you got a dip or peak in your revenue on the KPI dashboard by means of a data drilldown
— Determine the most efficient marketing activity.
- Predictive and prescriptive analysis for anticipating and forecasting
This is the professional league of data analysis. Predictive analysis helps you understand what’s likely to happen if nothing changes. It shows you:
— how your sales will grow
— which leads are most likely to convert
— what risks you can decrease.
Example of a table for the Plan vs Forecast report based on predictive analysis data
Meanwhile, prescriptive analysis based on sophisticated software and technologies helps you model the whole business in situations like a market change or turning off a whole promotion channel. In practice, you not only get to check any marketing hypothesis but also:
— achieve top-level personalization
— optimize product lines.
All of these levels of data analysis demand special tools. But even the most complicated e-commerce data analysis starts with the simplest tool for collecting and tracking data. We’ll show you an example of how to set up and perform basic data analysis for an online business.
Setting up e-commerce data analysis from A to Z
a. Define goals and metrics
There are five main goals for any e-commerce project:
All of these goals are about creating and maintaining your faithful audience and stimulating purchases.
Acquisition is about attracting new visitors to your website so they can convert by doing things like subscribing to the newsletter. Once you’ve established communication with a website visitor, the path is open to a purchase. But after a purchase, you’re only halfway done.
Now you have to retain the customer for as long as possible so they buy from you again. Loyal customers can even bring you cheaper leads based on referrals on social media or product recommendations.
The last step in defining your goals is choosing the main metrics you’ll track. Depending on the analytics tool you choose, some of your metrics will be automatically calculated for you — like website visitors in Google Analytics. But don’t be afraid to calculate some metrics manually or with the help of additional tools.
Your two to five main metrics should be meaningful — better to pay attention to purchases and retention than to acquisitions and referrals. We’re not saying acquisitions and referrals aren’t important, but there’s always a risk of making them vanity metrics that lead you to grow numbers, not actual revenue.
b. Set up tracking tags, filters, and conversions
Commercial websites install tracking tags on their pages to track events and observe customer behavior. A good analytics system can track real-time events and provide ad-hoc reporting.
Be sure your website data is available as soon as an event happens on the page. This will allow you to make decisions based on actual and reliable data.
What are filters? Filters are features of analytics tools that help you keep your traffic data clear and adjustable. Filters save you from:
- bot data. Sometimes bots visit websites even more frequently than real visitors.
- unwanted data. For instance, you may want to exclude traffic from your employees.
- mixing all your data in one view. You can create separate filtered views for analyzing traffic divided into different segments or categories.
Set up your conversions and custom goals in your analytics system to see what customer journey is so important for you that you want to know how many people complete all steps up to the conversion or how and why they drop off along the way. Every analytics tool will offer you predefined goals, but typically they won’t work for your website.
Adjust your goals to your website. Create meaningful names for conversions and goals to make them understandable in datasets and reports.
This is the basic setup for any e-commerce website. It’s the starter pack for tracking and collecting basic data from your website. At this step, you can also start reporting and dashboarding experiments.
Pay attention to details and keep in mind that your website is not alone in the universe and that customers are coming from places both online and offline that you can track and explore. So what should you do next?
c. Integrate with advertising accounts, CRM systems, call center software, and offline stores
This is the stage where you start to see that your analytics tool won’t be totally useful without your cost, advertising, and customer data — which can’t be collected by your analytics tool. The moment you wonder how to see true revenue data in your purchases report is when the real e-commerce data analytics starts.
There are two problems you’ll encounter at this point:
- The data you need is stored in different formats and sources. Maybe another organization even holds this data for you (for instance, if you outsource advertising services).
- You can’t merge your data without conflicts and all data needs to be cleaned. Dates may not fit the data schema, fields might represent the same thing but have different names, etc.
You’ll need to get integration tools and a place to store all your merged data. And you’ll need analysts to put everything in order. Your company might have some specific security requirements or a bias for certain analytics tools, so it’s totally up to you.
At any rate, you’ll get a whole monstrous ocean of data! One of the most interesting and overwhelming experiences at this stage is making your analytics help you not only increase revenue but decrease inefficient expenditures. It’s all about building an attribution model.
c. Build an attribution model and automate your marketing
The main aim of attribution modeling is to show what channels bring you the most revenue.
That’s an awesome feature because you can use this information to redistribute your advertising budgets (and even do so automatically) to make your efficient campaigns even more beneficial and stop all of your inefficient campaigns. You can get acquainted with existing attribution models or create your own.
Example of a funnel-based attribution modeling report
At this stage, you’ll need someone who specializes in big data. By working with them, you can find out how to fully capitalize on the data you collect to improve your advertising campaigns and multiply your total revenue.
Also, this is the best time to build some cool dashboards — one for all the main metrics you want to track daily or even hourly, one for your monthly reports, one for forecasting your yearly sales plan performance, and one for attribution results. These dashboards can’t even be compared with those you were able to build at the first stage of developing your analytics system.
Sometimes, analyzing e-commerce data makes your colleagues’ lives complicated. It’s also true that the first fruit won’t come fast. But the biggest risk is still avoiding analytics, not trying to adopt it.
No matter how hard the way, you should tackle all the tracking, collecting, merging, cleaning, and reporting issues to become not only a successful e-commerce business but a data-driven one.
This post was submitted by a TNS experts. Check out our Contributor page for details about how you can share your ideas on digital marketing, SEO, social media, growth hacking and content marketing with our audience.