Delivering optimal online performance for SaaS
Adoption of Software-as-a-Service (SaaS), which delivers critical business applications over the Internet, is skyrocketing. IDC predicts that SaaS will grow at a 25 percent compound annual rate through 2014. Economic and operational factors are driving this popularity. The SaaS pay-as-you-go business model means low up-front costs. SaaS firms offload hardware management, maintenance, security and upgrade tasks, reducing the workload within customers’ IT departments. At the same time, SaaS solutions promise a high-quality user experience, excellent performance and error-free operation.
How can organizations that offer SaaS applications assure measurable and repeatable service level agreements (SLAs) across the web and prevent user frustration, complaints, disputes, and ultimately lost business? The key is to understand the performance delivered to actual end-users, in real-time. A real-time view of the totality of delivered end-user transactions is the only means available to measure performance, availability and traffic load, and gauge satisfaction.
The challenges of delivering SaaS applications
SaaS firms face numerous challenges in order to provide better, faster, and more reliable applications.
- Business managers of SaaS firms can’t measure service levels accurately. Service Level Agreements ensure that applications perform as promised – but only if they are measured by actual end-user performance. End-user SLAs are critical to maintaining and increasing revenue as they directly affect customer loyalty and are imperative come contract renewal time. When a dispute arises over outages or missed SLAs, the customer and the SaaS provider need accurate data regarding SLA adherence so that discussions focus on facts rather than anecdotes about perceived user problems. However, information about SLA adherence is often unavailable because traditional management systems don’t accurately measure real-users.
- Problems take too long to find and fix. Often SaaS firms are able to detect errors only after users complain. Moreover, users often complain to their own IT department, which means that the SaaS providers are only aware of a small percentage of the errors their end users receive. This makes it virtually impossible to find and fix problems in a timely manner.
- Recreating problems is difficult. When problems are reported, it is difficult to recreate the issue and task the right technical team with fixing the problem.
- SaaS operators don’t always see the impact of a change to the application. Application development cycles happen quickly in the world of the Web. Content changes and application adjustments can all cause problems that impact end users. The difficulty is exacerbated by the desire to move quickly. Yet, even with pre-deployment testing, the impact of changes on end-users may be significant. It’s challenging to ensure problem-free operation of new functionality in the real world.
Real-time, end-user visibility tools answer those challenges and make it possible to efficiently deliver high-quality web applications.
The inadequacies of existing tools
SaaS providers have traditionally used several types of tools to measure the performance of SaaS applications, but these solutions don’t monitor from the end-user perspective, a significant limitation. These tools have value, but may not be relied on by themselves as they are unable to bring the end-user performance into focus.
- Device monitoring. Device monitoring tools create log files of data that measures the health of the systems, such as CPU and RAM usage. However, correlating data from these log files is a time-consuming process that makes it difficult to draw conclusions.
- Load generation. Used prior to deployment, load-generation tools deluge applications with requests in order to stress-test the application and identify problems that weren’t caught earlier in the development cycle. But the tests may not give an accurate picture of how the application will react when used by real users, who tend to act in an unpredictable manner, and when peak traffic demands occur.
- Synthetic testing. Synthetic testing simulates user transactions by frequently running scripts and tracking the success or failure of the script. However, these tests are only as good as the scripts that were written, which can never fully anticipate all problems. Additionally, these tests measure only a limited number of high-value pages, they generate significant additional performance-robbing traffic to the site and they are often too expensive and too infrequent to provide an accurate picture of performance.
The real-time, end-user perspective addresses these challenges
End-user monitoring and management tools address these challenges by allowing SaaS providers to monitor actual end-user transactions to measure performance, availability and traffic load. Thus, these tools inform SaaS providers of the service level of application delivery to every user in real-time.
End-user management tools also offer rich reporting and analysis on an entire application, customer group, or business process as well as allowing drill-down to individual users to aid with problem alerting, investigation and resolution. The result is faster detection and localization of incidents; an understanding of the impact of changes on customer groups and users; and accurate, authoritative measurement of service quality across customers.
End-user management tools have many uses, but they ultimately serve three important functions:
- Reporting on service quality for an individual user or group
- Measuring the impact of a change on real users
- Detecting, capturing, and localizing problems
Business benefits
By understanding the end-user experience, SaaS providers can also achieve substantial business benefits.
Better performance and fewer complaints means a SaaS provider can generate more revenue and retain customers. Customer satisfaction is the critical business values. When they are able to deliver promised service levels on a consistent basis and document service quality, SaaS firms see higher usage of SaaS services and have an easier time convincing customers to renew their licenses. When they can proactively detect and troubleshoot problems, SaaS firms can accurately determine the source of problems to fix them before they impact users and avoid finger pointing among the vendors involved in application delivery.
Conclusion
SaaS companies promise to deliver better performance for applications over the web than organizations can provide themselves. Achieving this goal requires that SaaS firms gain a better understanding of the actual user experience and establish a well thought-out plan for delivering and ensuring the most reliable service levels and user satisfaction possible.
Because the end-user view provides visibility into the actual, delivered end-user experience along with full forensic detail and sophisticated data collection, it enables SaaS firms to conquer the service level delivery challenge and furnish operational excellence, faster problem resolution, quicker rollouts and competitive differentiation. As a result, these firms can achieve greater usage of SaaS services and higher renewal rates.
biography
Ali Hedayati is president and COO of Coradiant. Ali
is experienced at building strategic partnerships and in developing ecosystems within new technology
marketplaces. Prior to his appointment as President, Ali led Coradiant's marketing.