Automation/ROI - how to calculate? (Part 2)
Task Automation
Automating IT tasks immediately increases the productivity and speeds-up delivery.
Instead of spending time freeing up server disk space, resetting passwords for Active Directory users, and restarting services, employees are freed to focus on higher-value work such as streamlining capacity configuration across a fleet of devices.
Note that the value realized from IT process and task automation is incremental. While automating a 20-minute task may seem insignificant, your perspective changes when you realize the task is performed ten times a day and there are hundreds of those tasks on a weekly basis.
How do I calculate task automation value?
The basic formula below will let you assign a monetary value to the freed-up time. This formula should be applied to each task you plan to automate so that you can add up the annual value per task.
TIME (spend on single manual task) x FREQUENCY (of performing task per month) x 12 MONTHS = YEARLY SAVINGS/VALUE FROM AUTOMATION
Assuming you spend 5 hours weekly on server maintenance such as disk space clean up. This is 20 monthly hours, at an average cost of $60 per hour, for a total of $1200 per month. On a yearly basis, this amounts to $14,400 in value.
Note: Even after a task is automated, it will always require some human attention in the form of confirmation or at times intervention (for more critical cases). Thus, you should reduce your estimated annual value by 10-15% to be on conservative side. In our case, final value would be between $1.4K - $2.1K.
Downtime reduction
The cost of system downtime will vary per business area. Downtime costs depend on two primary factors: productivity losses and business losses. Focus on repetitive incidents and disruptions for a quick win.
How IT Process Automation Helps Reduce Downtime
1. Preventive Maintenance - scheduled processes that run regularly such as cleaning up server hard drives, updating critical patches, checking backup logs to maintain the systems’ health, etc.
2. Auto remediation - automatically trigger workflow processes to handle critical events by performing corrective actions that restore service back to normal as fast as possible, such as automated service restarts on servers
3. Automate complex tasks - automated processes to monitor mission critical application across systems and networks and perform maintenance, preventive actions, etc.
To calculate downtime reduction value requires using a two-step process.
First, you calculate the yearly costs of a service or system outage – before any automaton.
HOURLY LOSS (of productivity or revenue) x AVERAGE DURATION (of system downtime in hours) x OUTAGES (number per year) = YEARLY DOWNTIME COSTS
Hourly loss. Estimate the cost for an hour of downtime of your critical systems by collaborating with the business partners to come up with a realistic estimate. Remember to include both loss of revenues and loss of productivity. For example, if SAP system is not available for 3 hours, what is the estimated loss to the business in terms of lost personnel hours (productivity)?
Average duration. Try to estimate the average duration of a system downtime event, based on previous system downtime information, such as service disruption events or incidents.
Outages. Estimate the average number of outages per year based on number of incidents or service disruption events.
YEARLY DOWNTIME COSTS x OUTAGE REDUCTION % (due to automation) = YEARLY AUTOMATION VALUE/SAVINGS
With automation we can estimate that we will reduce outages by 10-15% a year. If the yearly cost of outages is ~$500K, then automation would bring the value of $50K-$75K per year.
MTTR – Mean Time to Resolution
MTTR measures how long, on average, it takes to restore a system to its operational state after a failure. It is commonly argued that at least 50% of MTTR is spent in diagnosis, that is finding the root-cause of a problem. However, a large chunk of system failures does not require extensive investigation and can be identified and resolved with pre-defined actions. Focus on repetitive incidents and disruptions for a quick win.
How IT Process Automation Helps Reduce MTTR?
1. Alert Capturing - process workflows can be triggered automatically upon detection of system alerts using systems such as Netcool.
2. Troubleshooting - automated diagnosis processes can reduce the time to identify the problem and react.
3. Notifications & escalation - notifications can be automatically sent to relevant personnel. If no response is received, an escalation process can be triggered such as paging out next person on the team or team's lead.
4. Remote action - the owner of a problem initiates pre-configured workflow processes that automatically remediates a problem.
To calculate MTTR reduction value requires using a two-step process. You begin by finding the total yearly cost of incidents.
INCIDENTS (# of disruptions per year) x TIME (hours spent on resolving each disruption) x COST (of personnel per hour) x 12 MONTHS = YEARLY COST OF INCIDENTS/DISRUPTIONS
Incidents: the number of incidents and disruptions per month. Since incidents differ in their severity & priority, you can use this formula each time for different incident priority, like critical vs limited. Each will have its own resolution time and associated cost.
Time: the total time it took to handle a single incident or disruption.
Cost: the cost of support personnel that spent time remediating the problem. You may need to consider several cost levels for support personnel or use a flat hourly rate across all support levels.
Next, your yearly value of automation is calculated by using an estimated percentage of reduction in resolution time that automation delivers.
YEARLY COST OF INCIDENTS/DISRUPTIONS x REDUCTION OF TIME TO RESOLVE (in %, due to automation) = YEARLY AUTOMATION VALUE
Let’s use a conservative number between 50% - 70% as our reduced time to resolve. Thus, if the yearly cost of incidents is ~$400K, then automation would bring value of $200K- $280K per year.
Servant Leader | Process, Control, and Operations Automation | Risk Management | Process Improvement | Data & Analytics | End-to-end Vision | Innovative Thinker | Executive Communication and Reporting
7moValuable article and fantastic for actuals. Let's say you have an idea of how many different tasks you believe you can automate, but you're not certain yet how much of the task you can automate (i.e. full or partial automation). How much of a buffer (or range) would you build into providing a potential savings (time and/or money) as a result of automation? Example: You have 10 tasks you believe to be high potential for automation. Each task takes 10 hours a month to execute. You want to provide leadership "potential" time savings if you automate all 10. Would you do a range of potential time savings not knowing if you can fully automate each task? How much of a buffer would you build in? +/- XX%? - 10 * 10 * 12=1,200 hours - Potential Time Savings (10% buffer): 1080-1200 hours? You wouldn't do a +10% (1320 hours) because you can't automate fully and save more hours than it takes, right?