Overcoming the Second Biggest Challenge in the Cloud Space
As can be seen from the research below, managing cloud spend is the second biggest challenge in the cloud space.
Why is that?
How IT costs are managed within organisations has changed. Traditional procurement methods and centralised financial management are being side-lined when it comes to procuring new cloud services. Not least because cloud services promise to deliver faster delivery from IT teams – something businesses have been yearning for for years.
There is no longer the need to wait weeks to procure infrastructure, or install your data centres. All of this is available at the flick of a switch and can then grow as required. It’s great news and for many, this is made possible by simply having a payment approach lodged with the appropriate vendors.
But great news can quickly spiral into sticky situations if taken on face value. Let’s take a look at the challenges adopting a lighter governance approach can present:
• Potential for uncontrollable spend – any developer with the correct API and a company credit card can incur costs.
• Lack of understanding of what business services are driving the costs and whether the activity is necessary and delivering value.
• Issue with large bills for incurred services with no warning of the costs being incurred as they weren’t picked up until the bills came in.
• Lack of ability by finance to translate the bill, forecast the spend and track the budget.
Is this a new problem?
Well the answer to this is yes and no. Although public cloud is a new cost for many companies, the problem is similar to existing services. The good news is, we can draw knowledge and experience to help with addressing the problem. Let’s take a look at a few examples from the archives:
1990 IT spend
In the 1990’s, when IT spend was being challenged from all angles and justifications were needed to progress with any initiatives, strong approaches were developed to deal with this. Many companies developed Capacity and Performance management teams whose main drivers were to deliver an optimised service delivering required performance for the lowest cost point. The services analysed poor performance, forecasted IT demand and associated costs and provided chargeback costs to the relevant business units. This was generally a monthly process with forecast spend going out into next few years. The aim was to determine when the next major purchase of IT Equipment was needed to allow for the planning and budgeting of a significant Capex spend.
Whilst the 1990’s approach would work with cloud, the traditional monthly reporting cycle would need to be brought to near real time to deal with the exposure potential for runaway cloud costs.
Mobile Phone Bills
This the most common analogy used to explain the issue with cloud costs. You buy a contract with a mobile phone company and register a payment method. You have limits to your account, number of minutes, number of texts and a set amount of data, and yet, it is still possible to exceed these limits with an associated higher charge.
These higher charges were initially met with shock, but now processes have been put in place to address them. Hard limits on usage, text alerts when you are approaching a limit, unlimited resources at a fixed charge and expense management systems that can monitor usage across a company to get best value from the contracts. All these approaches can be applied against cloud.
Enterprise Software Agreement
When procuring new software became an onerous task, agreements were entered into to all drawdown of software products. These agreements would allow for a fixed limit to be drawn down either via cost of license numbers (similar to the limits of private cloud) or an unlimited deal with “true ups” at regular intervals to manage the costs and invoice for additional usage. Both these scenarios need careful management to avoid huge bills and blowing the available budgets on unnecessary software.
A whole specialist field was generated around software asset management, understanding consumption vs entitlement to ensure that spend was closely managed. As with the mobile phone and capacity management scenarios, there is a need for automation and management software to stay on top of all the moving parts to ensure that the correct level of management reporting is available.
So how does public cloud consumption differ from the examples above?
Well, for one, the brains behind cloud technology have taken on board lessons learnt from history and that means that today, it is far easier to get the relevant billing information you need to manage and analyse any potential problems. Problems such as:
* Public cloud is a real time, on-demand consumption model with potentially no limits or alerts on spend unless you want to introduce them. In finance terms, this makes it more of an Opex purchasing model rather than Capex.
* In some of the approaches above there was a potential for human error in the charging process, whereas in cloud billing it’s all automated. The bill you receive is itemised and it is very easy to tell what costs are attributed to what activity.
* Cloud billing is designed to provide meta data about who/what is consuming the services to help with the management. In the traditional capacity planning model of the 90s, this was one of the hardest things to achieve; true cost transparency. It led to the development of special toolsets and functions to add to applications to understand what is needed to deliver business services and the importance of the Configuration Management Database (CMBD). All this functionality is built into cloud and just needs to be exploited.
* Understanding billing without a reasonable understanding of cloud services is sometimes difficult, especially when it comes to deciphering the contents of the cloud bill or when there is a significant number of lines on the bill adding to the complexity.
Six steps to optimising your cloud cost management
Given the above potential problems and availability of skill sets already in your organisation, this is an approach we would recommend when approaching your own cloud cost management:
1. Understand Cloud Account Structure – gather together all the accounts used by the organisation. Review your company credit card billing and expenses to bring together all cloud spend.
2. Identify a toolset to help you with the task in hand. Cloud providers can provide native tooling such as Cost Explorer in AWS, Azure advisor from Microsoft and Google cloud console, or you can invest in a multi cloud toolset such as Cloudability.
3. Build a dashboard view of all your cloud costs to allow you to compare and contrast usage and understand consumption vs budget. This is a useful way to educate your teams in the use of cloud and impact on costs. Build some automated alerting to pick up on any anomalies in spend that can be addressed immediately rather when the bill comes in.
4. Map the costs to consumers through use of cost transparency exercises and cloud provider functionality like tagging or labelling. Understand breakdown of spend across business applications and production vs non-production environments.
5. Review current cloud consumption and make sure that idle resources are turned off or consolidated. Undertaking a right sizing exercise to ensure optimal use of all the resources. This will give you a baseline consumption model.
6. Review your cloud purchasing model to maximise use of cloud discounts and savings plans. This will significantly reduce your cloud spend and introduce an element of capex spend into your finances.
Whilst the above will help provide the transparency needed to truly manage and analyse your own cloud cost, groundwork will need to be put in place to ensure you’re regularly getting the insights you need. Join ECS for a joint webinar with Apptio on July 16th to discover six simple steps to staying in control of your cloud spend. Register your space here.
To discover more about how you can establish a Cloud Centre of Excellence and operating model to deliver and run cloud within your wider organisation, click here.