What is APM?
APM got its start in the academic research community, where early efforts were focused on software quality improvements, such as reducing excess code loops or tightening up performance and usability issues.In the past several years, the area has grown to include all sorts of business insights that can be gleaned from watching how end users consume software. It isn’t just about preventing downtime, but how a business can become more profitable and increase revenues. Now there are APM products that tie together web performance metrics, business optimization methods, and other techniques.
The growing importance of APM
This expansion of APM has happened as a result of the collusion of several factors that have made it a more critical tool in the business software war chest.
- Cloud computing has become easier and cheaper. The days where you had to build your own servers and data centers are over. Ralston Foods is probably the largest cloud-only company without their own data center, and their revenues are in the billions. Thanks to the numerous cloud providers such as Amazon, Rackspace, and others, it is very inexpensive to store loads of data online, where analytics programs can quickly access it. The cloud also makes it faster to bring up analytics software: previous on-premises versions were months long consulting projects.
- eCommerce has become ubiquitous. Many bricks and mortar companies have become more oriented towards eCommerce and online sales just because there are more customers who want to purchase goods and services online. And the more companies go online or move more of their business to online sales, the more they need analytics to understand these purchase patterns and customer behaviors. Even for the most traditional “rust belt” type of enterprise, APM can enrich their understanding of their business and give them information about their customers in ways that they might not see with traditional metrics.
- IT is becoming more comfortable with Big Data techniques. What once was the province of the largest Internet companies such as Yahoo, Google, Facebook and Twitter is now available and more commonplace for many companies. You can now find Hadoop Big Data ecosystems in many staid and more traditional Fortune 500 companies. As an example, the IT staffers that brought up Hadoop for the giant retailer Sears eventually spun off a side business in helping others get started with the technology.
- Web analytics and BI tools have paved the way for better analysis tools. There are plenty of IT managers that are accustomed to looking at the analytics for their website interactions, such as daily page views and visitor profiles. There are also plenty of business intelligence tools that are available too. Both have made using analytics more approachable and acceptable now.
- The web stack has gotten too complex for manual methods. It used to be that a single developer could bring up a website with a little HTML code and a few ordinary PCs. Those days are pretty much over. The web technology stack now includes dozens of different servers and services and it is rare that a single developer can have a handle on the entire entity. This means having appropriate analytics and automated tools that can help measure the business implications is more important. Now software development teams can instrument their code and observe the results as they go from build to build.
- Customer experience is king. One way that businesses can differentiate themselves is by paying attention to their customers. This isn’t anything new: Nordstrom’s department stores have been doing this for decades. But what is new is being able to tie the customer experience into some solid metrics that can be quickly acted upon. One way towards improving customer experience would be to have analytics drive the business logic of a website.
The future of APM
These are exciting times for APM. The future holds a lot of promise. Here are some highlights of where APM could go in the next few years:
- The amazing view of customer behavior. With APM, you can gain a big competitive advantage since you have a deep understanding of your customers, what they buy, and why they buy.
- Software as a team sport. And the team is getting bigger as APM makes software more approachable beyond the small cadre of programmers that can code. As Ahmed Hassan wrote in the IEEE Software special issue on “What’s next in software analytics,” APM “needs to service a project's various stakeholders-its marketing, sales, support, and legal teams-not just developers. “ APM can help prove that software can be a cost-effective mechanism to increasing profits and cutting costs.
- Proving its relevance to everyone. The alternative is doing nothing, or having incomplete information. “It is in proving the cost effectiveness of our techniques and addressing the need for tools and techniques that leverage our knowledge of software engineering,” writes Abram Hindle in that same IEEE issue.
- Dashboards and better data visualizations everywhere. With APM, you can have a plethora of simplified dashboards that can deliver data needed for complex business decisions on smartphones/tablets.
- The democratization of business analytics. All companies run on software, and with SaaS, they can all get sophisticated, useful access to deep data on how their software, and their business, is performing so they can make better/faster decisions.
- Better data mining techniques. The past data mining techniques have not kept up with how software code can be instrumented and linked to a number of factors, customer behavior, and infrastructure status information. For APM to shine these links have to be made easier to implement.
- Better pairing and use of IT analysis with programmers. For the fans of the show “24,” Sunghum Kim writes in the same IEEE article that we need more Chloes and fewer Jack Bauers, where the analysts (the Chloes) can provide better assistance to the developers and guide them to make the right decisions.
