What is a KPI and why should you care?
Any relevant business has and tracks a KPI, if not multiple.
In pretty much every article I write, I talk about how to design your cloud architecture to help your business meet your KPIs. If you don’t know what a KPI is then that likely doesn’t make much sense to you.
A Key Performance Indicator (KPI) measures performance, as you would expect, but can also measure agility, efficiency, and other traits that are important to your business. Typically they relate to critical success factors for your company as it relates to ongoing operations. In theory, the best way to measure up your business operations to see if you are meeting your KPIs is to track everything that happens over the course of a given day. But is that possible? I would say yes, depending on how specific your KPI is and how much you leverage cloud computing to help you with it.
Lets take a couple examples of KPIs from well established companies so we can get a better idea of what one should actually look like. For this example I will use DoorDash, Lyft, and AirBNB.
DoorDash is a food delivery service that gets food from the kitchen of a restaurant, to the front door of a consumer. With that in mind, we can safely assume a KPI for them is cost per delivery. If they are spending more money on a delivery than what they would actually profit from a delivery, something needs to change or they would not be in business for much longer.
Lyft is a ride share service that takes customers from point A to point B. Since their business is based around ride sharing, we can say a KPI for Lyft would be cost per ride.
AirBNB allows people to open up a house or some spare room in their current house to rent out to various travelers. Their business revolves around connecting property owners to travelers and providing a booking service to take away all the management and hassle of doing that from the property owner. So, we can say for them a KPI would be cost per booking.
Hopefully now you have a better understanding of what a KPI is and why any business worth its weight in salt would have them and actively monitor them. However I did mention earlier that in theory, the best way to see if you are meeting your KPIs is to track everything. But is this feasible, or even possible? There are certainly many approaches you could take to measure your operations, however if you utilize cloud computing you can make this exponentially more easier, efficient, and accurate.
The most straightforward way to track your operations and measure your KPIs within your cloud operations is by using tags. These are often ignored when setting up resources as they are considered optional. However if you want to fully optimize and control your costs you need to tag everything possible. The reason why is simple:
Say for example you have a 3 tiered web application with a few servers running in the public facing front end, a couple of private database servers on the backend, and some application logic servers in the middle doing processing and business logic. When your bill from AWS comes around if you do not have any tags, you are only going to see what you owe by service (EC2, RDS, ELB etc.). So there is really no way to know which of your database or front end servers are being utilized the most, therefore it would be hard to go in and restructure your architecture to be more cost efficient if you are pushing your budget.
On the other hand, if you were to tag each and every one of your resources (tags being user defined metadata consisting of a key-value pair) with some naming conventions like environment, role, owner, or purpose, you could easily separate out your resources into different categories. Then you can use a service like AWS cost explorer to specifically see which of your resources are costing what amount and readjust from there. Here is a link to some AWS documentation on tagging, of course other cloud providers will have their own version of this that you can look into as well.
The better and more consistent you are with tagging, the easier it will be to track your operational costs. And when you are able to track cost, along with revenue, it becomes very clear how you are measuring up to your company defined KPI. And if you are not doing very well then tagging will allow you to restructure some things so that moving forward you come closer to meeting or exceeding your KPIs.
Hopefully you were able to find some value from this, and see yet another reason why cloud computing is so valuable and integral to any and all business operations. If you have used tags before and have some best practices you would like to share I would love to hear about it.
As always, best of luck to you on your continued journey through cloud computing.
