Blog
Do you know the ROAO of your application landscape?
Have you ever investigated how (in)efficient your application landscape is? Aside from the irritation it's causing your employees ... you're probably leaving a lot of money on the table by not optimizing your applications. Bold statement? Maybe ... Mostly 25+ years experience of observing behind the scenes at organizations and all that lost money of missed application optimization.
Application optimization may sound like a technical term, but in essence it's something really simple: improving the digital tools we use every day. The goal? A supple, fast and less frustrating work experience for everyone in the organization. And ... saving cost. Because that last point ... gets way too little attention.
In this blog I take a deep dive into the world of application optimization. I investigate how it can reduce the daily irritations of employees, raise productivity and how it can even contribute to a better work-life balance. The financial advantages are big and a lot the gain is in an improved user experience and employee satisfaction.
Invoice stream costs extra FTE's
Example: At one of our customers a digital invoice stream has a manual check at the end. When the application was rolled out the invoices displayed quickly, but over time it started taking 60 seconds between requesting an invoice and being able to check the invoice. Users got used to this automatically. We investigated en discovered the cause of the 60 second waiting time. After proving insight into the waiting time we could fix the causes and quickly reduce the waiting time to 20 seconds. The employees responsible for checking invoices were delivered from an irritation. For every 100 invoices 4000 seconds of worktime were saved, that's over an hour, for every 800 invoices daily an employee - that's an entire FTE - could be reassigned to different work.
Logging in to BI environment takes minutes
Another example from our broad experience: A BI environment on a data warehouse takes minutes to let users log in. They can literally get coffee (and they do), drink it, and just as they're considering getting their next cup of Joe, finally a menu with reports appears. Every 2 hours there's a 10 minute period that it's even slower, twice as slow than it is during the rest of the day. The irritation with the users is easy to imagine. With 250 users every 2 minutes waiting time means that 1 person, 1 FTE, gets paid an entire workday for ... waiting. Said differently, for every 2 minutes gained you need one employee (a full FTE) less. That's serious money.
Analysis of the environment by thee employees of Sciante revealed 3 causes of the delays:
-
At login time the BI software tries to cache a large part of the database, to be fast later on. A lot of data is fetched that the employee never needs in his reports. That extra data only delays logging in.
-
The virus scanner scant all data retrieved from the data warehouse. Scanning data is not required as the data is not executable code. Virus scanning data makes fetching that data dozens of times slower.
-
Every 2 hours a data refresh is performed that takes about 10 minutes. Refreshing of data has such a big impact on the data warehouse that it's only half as fast.
Resolving those bottlenecks speeds up logging in to 13 seconds. It also speeds up the average render time of reports from 20 to 2 seconds. Irritation solved. Financial gain is a half hour per employee per day. Every 2 minutes is an FTE, so 30 minutes that saves 15 FTE total. On this BI environment alone.
It is not normal
When inefficiencies continue long enough we get used to them and we start considering them normal. But waiting for IT is not normal. Not necessary either, and certainly undesirable. If Google can spit out pages of search results in less than a second and if ChatGPT and Claude can produce entire volumes in seconds, why would it be normal if your application takes 10, 20 or even 60 seconds to produce a report or an invoice? I say it often and I'll say it again: software wears out. And if you don't maintain it that wear will cause you trouble. Google, OpenAI and Anthropic maintain their software and fix the wear. They measure how fast their applications are and take action when the speed deteriorates. And not just by throwing hardware at it limitlessness, but by optimizing their software. And don't go thinking, "yeah listen Hugo, efficiency at their scale is important, but we're much smaller". With them it may be the difference between profit and loss, living or dying. For the somewhat smaller organizations it can make the difference between staying ahead of the competition or trailing behind them. Seems like a big win to me.
The route to efficiency
If you have a map or a floor plan, knowing where you want to go is not enough. You also need to know where you are now. If you don't you can't figure out the required route. Application optimization, like anything else, starts with figuring out where you are now.
Start by determining which applications are in your landscape, which ones are critical for your organization and how many employees use them. Then you measure how long employees need to wait and what they're waiting for. That tells you if you have a problem and how big that problem really is. You can also calculate how big the expected profit is, in reduced irritation and saved cost. We can give you that insight in 1-4 hours per application.
The next step is finding the causes of the delays and the solutions for them. You do that by performing (technical) measurements on problematic applications. That can be done real quickly. When done right you can know the largest pain points in 1-2 weeks and also know how to fix them. We have a special service for that, that finds these pain points with no fuss lightning fast. You don't need any IT performance knowledge and you don't need to interpret cryptic dashboards. You do get actionable insights you can apply to your application quickly as a result.
The last step is of course implementing the fixes. If it's an application you built in house, you put your people to work. If it's a bought or SAAS application, you'll often, but not always, need the supplier.
This gives you a high return on application optimization (ROAO).
Keep the irritations at bay
If the application that worked brilliantly after being rolled out, is annoying for your employees now, then it's time for application optimization. If you create a one off solution now, you have a serious chance of returning problems in the near future. And that's the last thing you want. You want to secure your investment in optimization by monitoring the improved application continuously, or at least monthly, for efficiency. That takes much less effort than you might think. Take a base measurement right after the improvement and keep an eye on if the application is improving or deteriorating - automated of course. If the application is degrading measurably, but not yet noticibly, your users are not affected yet and you can can take action before a grumbling murmur becomes discernable in your organization.
We'd love to help you
Today's application landscapes and the underlying technology is complicated. I'd love to help you take that hurdle step by step. That start with a 15 minute Zoom call with me, no strings attached. In that meeting you'll find out where you are and where you want to go. And if you want our help with the execution and the followup.