Beginner’s Guide to APM

July 3, 2016 by Leah Na'aman

Website Performance is a critical factor in determining online business success. Whether site problems occur with a network or server, or product images fail to load properly, performance issues like these drive the same potential result – frustrated shoppers abandoning your online store. And unfortunately, customers don’t forget, so once they’ve had a poor experience it’s highly unlikely they’ll return. This makes giving shoppers a smooth in-store experience an absolute must for any online retailer. Now, perhaps you’re already monitoring your site’s performance. Maybe you’re using analytics tools to identify behavior patterns that point to performance problems.  Maybe you’ve tried writing your own scripts to pick up on issues in real-time. Or maybe you’re using out of the box monitoring tools to track performance over time. But if not, or if website performance monitoring is new to you, this introductory guide to APM (Application Performance Management) will help get you familiar with the must-know concepts, so you can take control of your site performance and acquire more happy shoppers.

So, what is performance monitoring, really?

The monitoring industry has often complicated the definition of APM throughout its history. And has always referred to the systems, tools and methodologies used to monitor performance, and the many metrics they address, in an attempt to define it. But at Shoppimon, we think it’s much simpler. We view APM as technology that helps site owners and operators answer a basic question, “Is my business really open for business?” And just as importantly, “If it isn’t – why?”

So how exactly does APM help e-stores get to the bottom of these questions? By monitoring their performance under real or simulated conditions, or both, and alerting teams to any problems that are encountered.

A few of the topics APM helps address are:

  • Site speed
  • Downtime
  • Errors
  • Performance trends
  • The business impact caused by store issues

You’ve built an incredible shop, so how much of an impact do performance issues really have?

You’ve invested in great UI-focused design, you have a strong development team behind you and a great product to offer you customers. With all of that in mind, performance can seem like just another thing to worry about. But the fact is, things can and do go wrong with Magento stores, no matter how many resources we’ve put into them. In fact, 100% of Magento stores will suffer from an issue in a given week. So because you’ve invested so much in your store, understanding and controlling its performance is an absolute must.

Let’s first take a look at a few of the metrics that have a big impact directly on performance, or can signify that there is something wrong with a page or site. There’s a good chance you’re familiar with some of them from your sales and marketing efforts, but let’s see how they relate to performance:

Speed– This is hands down, one of the most critical KPIs to watch for any online business. Did you know that 40% of shoppers will abandon a website that takes more than 3 seconds to load? In fact, Google includes site speed as one of the factors that impact your search ranking, meaning that if your site’s slow, no one will even be able to find you. So not only is site speed a big part of the overall store experience for your customers, but it can have a dramatic effect on the number of visits you see and subsequent sales.  

DowntimeDowntime is arguably the most immediately damaging to your bottom line. And at Shoppimon, we monitor two different types of downtime, Server and Business. Server downtime is where a site is down and completely unreachable, and business downtime is when shoppers are able to reach a store, but cannot complete a purchase. But no matter the type, whenever downtime occurs, you not only stand to lose first-time buyers, but with prolonged or regular outages, you risk loyal customers as well. At Shoppimon, we estimate that stores in the ecommerce industry are losing $142 billion every year due to performance issues that cause downtime.

Repeat PurchasesAs consumers, if we’ve heard great things or truly need a particular product, we may push through a store experience with a few bumps in the road, but we’re far less likely to do it a second time. With so many online stores to choose from today, it’s very easy for shoppers to look elsewhere when performance issues occur. So, low repeat purchases can signify a real performance problem.

Cart Abandonment– There is a long list of reasons that nearly 70% of shoppers abandon their carts without completing a purchase, many of which boil down to the psychology behind individual decisions. However, If you’ve recently noticed a spike in cart abandonment, your online store’s performance might be behind it. Anything from broken checkout processes to payment security concerns will directly impact your cart abandonment rate.

Bounce Rate Changes in bounce rates can signify many different issues, however they can also point to performance problems. If shoppers arrive to particular pages on your site and encounter error messages, lengthy load times, missing images, buttons or fonts, they’re not going to click-through to more pages. Issues like these risk both organic and paid traffic, and stand to pack a real punch to your bottom line.

Ways to monitor an eCommerce store

There are many strategies, methodologies and tools out there for site monitoring, each with their own strengths and challenges. Below we take a closer at some of the available options, how they work and their limitations.

Real-User Monitoring

How it works:

Real-user monitoring captures performance metrics and trends under the pressure of real users surfing a site. By monitoring actual performance data from the experiences of store customers, it stands to provide an incredibly accurate picture of how your site is performing for your customers, and alert on real issues as they occur.


  • Real-User monitoring is dependant upon actual site usage. Therefore, if not enough shoppers are browsing, searching, viewing and completing purchases within a given time period, you may not be able to get an accurate picture how your site will perform during peak shopping times, and errors or slowdowns may go undiscovered until it’s too late.
  • Because this type of monitoring only tests real usage, problems cannot be identified until they have already occurred, meaning sales and customers may already have been lost.

Synthetic Monitoring

How it works:

Synthetic monitoring tracks the same parameters as as real-user monitoring, but with traffic that is generated specifically for the purpose of testing, trending and reporting on site performance. Because there’s no need to wait for real users to visit an eCommerce store, you can potentially monitor stites either in production or before they are launched. Additionally, you can track performance under any set of conditions you choose. And because real traffic isn’t required, it’s possible to identify store issues long before they are experienced by real customers.


  • Because synthetic monitoring does not monitor store performance during real use, it will not identify issues when they are encountered by actual shoppers.

Server-side Monitoring

How it works:

Server-side monitoring requires an agent be added to a store’s code, or additional software to be installed, for it to report on exceptions that occur when running backend software. This type of monitoring focuses on the performance of various server-side components of an app, and will log issues and exceptions that occur and cause disturbances in a shopper’s in-store experience.


  • Because this type of monitoring is located on the server-side, it is not able to identify when and if issues occur on the client-side, allowing many store problems to potentially go unnoticed.
  • Server-side monitoring is always executed with an agent, meaning modification of a store’s code or installation of additional software is required to begin monitoring.

Client-side Monitoring

Client-side monitoring, unlike server-side, can be done either with or without an agent. This gives you the option of monitoring without the need to add code to your online store. Client-side monitoring can, however, refer to different use cases. Two of the most common definitions of client-side APM are either monitoring the performance of the client-side technology stack alone, or monitoring both the client-side technology stack and the touch-points between client and server-sides. Interestingly, while client-side monitoring can’t directly see the cause of issues that originate on a site’s backend, it can infer server-side problems based on client-side performance.   


  • Without an agent on the server-side of an application, it is difficult to identify the exact cause of backend performance issues.

A quick recap

To truly understand the performance of your store, get detailed performance metrics and trend performance over, time there is no one failsafe solution. The different forms of testing can compliment one another, but whichever method or methods you choose, the most important thing for you achieve with APM, is a deep understanding of store performance and immediate and contextual alerts on any problems that occur.

As online store performance has a dramatic effect on whether or not both new and repeat customers go on to make a purchase, having a firm grasp on it is a must. So whether you choose to monitor using synthetic or real user solutions, server-side or client-side, out of the box or in house scripting, be sure to take full advantage of your APM.