Calculating the Cost of Salesforce Integrations
by Alec Pinkham Alec Pinkham on

Salesforce is an integral part of many companies’ sales organizations. It’s a SaaS application that’s billed per head, but the selling point for businesses  is the time savings that sales reps can achieve with the organization and optimization of account data.

Alas, in a world of app integrations, time savings is not guaranteed, and a business-critical application like Salesforce comes with dependencies and limitations. But IT teams don't stop to actually measure the use and effectiveness of this tool very often.

At AppNeta we are always on the hunt for new integrations to add data to our sales processes. That added data can bring important details and perspectives for our customer teams. But we’ve also run into issues. We’ve heard complaints of Salesforce slowdowns. We’ve had complaints of single integrations that seem to be the problem. Quantifying the issue is difficult, unless you have AppNeta monitoring tools. We use the milestones function in our synthetic scripting engine to split out timings for our various integrations.

Testing Salesforce Integrations in Real Time

We selected just a few of these integrations to test out with our Experience monitoring function. We used the AppNeta synthetic scripting and milestones functions to identify the time it took to load the data from each integration within the Account view of Salesforce.

Let’s start with the baseline we established. The AppNeta script loads the Salesforce home page and then logs in to our instance of Salesforce. Each step takes around a second, which fits with expectations. Through normal web practices, the page will appear complete before the final load time that is shown here.

Monitoring Salesforce integrations with AppNeta

However, once we add the first few scripts, the time the Salesforce page takes to load dramatically increases to almost 10 seconds. (The purple diamond indicates a script change within our Experience monitoring.) If our sales team needs the data that’s within these integrations, we’re in trouble. Multiply a 10-second load time multiplied by the number of Salesforce accounts accessed per day (a conservative estimate is 100 accounts), and you’ll get a little more than 16 minutes of delay in a single day for every rep on the floor.

AppNeta Experience monitoring shows Salesforce integrations

When we add in the next set of integrations to the script, we observe another dramatic increase in the overall load time.

AppNeta Experience monitoring shows milestones

That initial load script had an average load time of 2.2 seconds. With these integrations, that load time is now 11 times longer, with an average load time of 24.4 seconds.

On average, each new integration adds 4.35 seconds of load time to the data your sales team needs. This is time added to every account load, and we are only looking at four integrations in this example.

All this lost time waiting for Salesforce translates into lost productivity and money. From the analysis above, we can assume 20 seconds of additional load time for each account. At 100 accounts visited per day (an admittedly conservative estimate), each rep would waste roughly 30 minutes waiting for data each day. At a sample $25 per hour per sales rep working 8-hour days, each rep on your team earns $200 per day, from which $12.50 is wasted Salesforce wait time. Over the course of a 50-week year (taking paid time off into consideration), each rep would accrue $3,125 in hourly pay just waiting for Salesforce integrations.

So with 100 sales reps on a team, your company is paying more than $300,000 for employee time just waiting for Salesforce integrations to load data. This is also a significant amount of time wasted, with 125 hours per rep per year, or a total of 12,500 hours across the entire 100-person team per year.

Check out this recent post on calculating the cost of slow applications for more on how they affect productivity and waste money.

Managing Salesforce Integrations

We’re certainly not suggesting that you avoid Salesforce integrations. Rather, it’s important to know what the user experience is so you can be effective in your troubleshooting, and plan ahead to avoid issues before they start. For example, much of the delay with integrations is not attributed to Salesforce itself, so calling Salesforce support would be a waste of time.

This problem is growing, too: Even a quick short Google Trend analysis shows the increasing interest in Salesforce integrations (with more than 14 million searches). As SaaS applications abound, and APIs open up more avenues for integration, it'll become even more important to understand how applications are working together.

It can be easy to see cloud and SaaS applications as somewhat mysterious or unknowable. But for the sake of your help desk ticket queue, your Salesforce users and your budget, find out what’s happening in these critical applications every day. Knowing these numbers lets IT improve slow applications and make end-user experience better—improving IT’s reputation in the process.

Filed Under: cloud computing, Performance Monitoring

Tags: API , SaaS applications , salesforce