The Performance Beacon

The web performance, analytics, and optimization blog

mPulse “What-If” dashboard: How does it work? And why does your business need it?

How SOASTA predictive web performance analytics works

How many times have you asked yourself “What if?”

“What if we could go to Mars?”

“What if we had autonomous vehicles?”

And how about something you could take action on right now, not years in the future? How about:

“What if my digital property had better performance? How would that affect the bottom line of my company?”

Not only can you ask this question, you can answer it with the help of mPulse real user monitoring (RUM).

mPulse is an integral component of SOASTA’s Digital Performance Management — a powerful platform that correlates performance to user experience and business metrics. To help with this correlation, mPulse includes a default dashboard entitled “What-If”, which uses predictive analytics to tell you the impact of performance changes on conversions, revenue, and other business metrics.

“What-If” scenarios

Intuitively, we can all understand that the slower the site, the higher the bounce rate. (Users are impatient, after all.) And the higher the bounce rate, the less time users spend on site — and hence the lower a site’s overall conversion rate. For e-tailers, conversion rate is one of the key business metrics that is an indicator of overall business health. (There are other user engagement and business metrics, of course, which can and should be correlated to performance and load times, but in this post, we’re going to focus on conversions and revenue.)

Below, you can see a screenshot of the default “What-If” dashboard view of an anonymized retail digital property.


Observe the histogram and the distribution of user sessions and conversion rates correlated on the same graph, as well as Conversion, Revenue, and Session Load Time metrics called out on the bottom with adjustable sliders.

Scenario 1: How do Conversion and Revenue change with better performance?

Let’s say your users are experiencing a 4.81 second average Session Load Time, and you would like to understand if expending resources into optimizing the user experience is going to yield positive ROI. All you have to do is drag the slider to the left to decrease load time to a manageable 4.50 seconds, for example. In the screenshot below, you can right away see the positive impact this has on conversions.

In this case, you could potentially increase revenue by more than $36K monthly. (Please note that the default timespan for this predictive dashboard is 30 days, but these time parameters are customizable.) Now you can be armed with this crucial data and make a clear pitch to make performance improvement a priority, with real ROI figures to support it.

Scenario 2: How do Conversion and Revenue change with degraded performance?

Now, conversely, let’s suppose your devs are telling you (or you, as a dev, are telling other people in your company) that the next release may impact performance negatively, and the average Session Load Time will go up to 5.19 seconds. Simply by moving the slider to the right, you can see in the screenshot below the potential negative impact to the bottom line of such a release.

In this case, conversion rate dropped by 0.07%, and revenue dropped by over $40K. This is critical data that you can use to ward off a potentially detrimental release in terms of performance.

Scenario 3: How much does Session Load Time need to improve to achieve an X improvement in revenue?

Lastly, let’s imagine that you want to achieve specific revenue targets or you have a fixed budget to improve performance. In this case, you can adjust the slider under Revenue to the projected number, and understand the targets that the dev team needs to achieve in terms of cutting down on Session Load time.

How does What-If analysis work?

So what’s the magic (aka math) behind this powerful dashboard?

First, mPulse uses a log-normal distribution to model sessions across load times. This is a model that creates the histogram in the graphics above and most closely resembles the true user behavior across a variety of websites. When the sliders in the “What-If” dashboards are adjusted, the model and histogram adjusts accordingly, and calculates a new conversion rate, from which revenue is derived as well.

In this case, revenue is simply a function of the change in conversions multiplied by the average order amount. Of course, it may and will vary depending on the absolute number of sessions, which may change over the course of a data set.

Conversion and revenue data sources

Conversion and order total (revenue) data can be automatically scraped by mPulse in order to make the “What-If” dashboard seeded and actionable. This is done by creating custom metrics in the Metrics tab of the Application configuration window of your property.

web performance predictive analytics

The source can be defined as a URL pattern of the Order Confirmation/Thank You page, for example, or using a JavaScript variable. Checking “Conversion Metric” or “Revenue Metric” on this form will send the data to the “What-If” dashboard.


There’s no magic in the “What-If” analysis dashboard. While the future can never be predicted with 100% accuracy, it can be modeled with a high degree of certainty using past performance data and sophisticated predictive analytics.

The more data available, the more accurate the prediction will be. This is why mPulse collects and stores all your website’s user data, forever.

The next “What if” question you may want to ask is, “What if we wanted to make the site faster? Which pages should we focus on first?” There’s an app for that, too. Using the Data Science Workbench module for mPulse to perform tasks such as Conversion Impact Scoring, you can take a deep dive into your user data and get answers to increasingly challenging questions.

Try mPulse for 14 days of free real user measurement

Edward Isarevich

About the Author

Edward Isarevich

Edward has an engineering background, and has developed his career both on the analytical and creative side. He has led teams and projects in product and marketing for software companies in various stages of development, including his own startup. As Senior Product Marketing Manager at SOASTA, Edward demonstrates to enterprises the value of digital performance management, deep data analysis, and the relationship between performance and business metrics.

Follow @isarevic