Success or failure of e-Commerce sites boil down to transactions. In brick and mortar stores, transactions are handled by point of sale systems. Their operation is outsourced, and their complexity is low. But for e-Commerce applications, the transactions not only rely on the processing API you choose, they also rely on the servers, the performance of the web application, the user experience and more. And your ability to adapt to the users in real-time, is what helps you drive a repeat visit and compete with other sites.
E-Commerce is no longer novel, and the competition is fierce. And for many sites, new users are going to demand the first-time experience is great. This is a first impression, and hopefully not your last. Users today demand efficient user interfaces, responsive sites, and stable backends that serve results quickly. In most cases even a small irritation results in a bounce.
Beyond just pleasing the users and assuring uptime in order to compete, you need to evolve your application rapidly; evolve to respond to user behavior, application performance, promotions, sudden changes in the environment, product placement, and calls to action. The only way to do this is with strong back-end and application analytics. Some of which you find in your standard web logs. But web log analysis only gets you part of the way.
So what is the core element that can help ensure transactions are successful?
In order to be responsive to changes in your infrastructure and application, you need to have all transactions and interactions with your application and backend collected and stored in one place. The collection is pretty straight forward; but the analysis is the hard part. This is where the data turns into real value. Often times we think of analysis as general periodic reporting. This will always be a part of it, but with the goals above, you require real-time. And any automated, immediate alerting on anomalies is critical.
Beyond the maintenance stuff, you also need to know enough to evolve the application. In order to evolve with your users, you need to understand their behavior as well. This involves the ability to correlate infrastructure data, application performance, and user interactions. Or, to understand the average page load time that your users tolerate you must correlate application and system data with each other. This type of data analysis helps you to make decisions on how to change and evolve your application to drive more transactions and happier repeat customers.
Here are the types of data you can leverage for an e-Commerce application.
- Server logs – Logs at the server level monitoring processes and performance.
- Web logs – Logs of all requests to and from the web server.
- Error logs – Logs that store data about failed requests (missing links), authentication failures, or timeout problems.
- Cookie logs – Logs to help boost the transaction-less state of web server interactions, enabling servers to trace client access across their hosted web content.
Focusing and analyzing this type of key information will tell you almost everything you need to know including: .
- Download timing: Helps you to understand how long it takes for a server to take the request and download a page for the user. In other words, how long your application makes your customer wait before he/she complete a transaction.
- Exit pages: Top Exit pages tell you which pages people visit immediately before they leave.
- Backend response: How quickly does the backend respond to user queries; and what are the shopping cart checkout processes.
- Peak load times: Peak load times let you know the general trend of load on your application. This is important because this is also the most risky times for long page loads, and possible outages. You can plan for scaling when you know this.
Traditional web analytics are not enough. Now that applications are measured as an entire stack not just code, you have to also analyze the data that each layer provides as a whole. This includes the ability to scale on-demand; to optimize workloads for faster transactions; and to monitor trends to avoid an outage.
Ultimately, your main objective is to create an e-Commerce site that is always available to accept orders, boost your sales, and increase end-user satisfaction.
Log Management and Analysis services are one of the only places you can store all your data for real-time analytics and correlation.