CAPACITY AND DEMAND
Embedded Files
Meet varying demand while keeping staffing costs under control with flexible scheduling.
Meet varying demand while keeping staffing costs under control with flexible scheduling.
1 | Mastering the challenge of capacity management
1 | Mastering the challenge of capacity management
Capacity and Demand Management (C&D) is applicable to most business operations and is especially critical to business at scale that have shifts in daily, weekly, or seasonal demand. Want to hit your SLAs? Want it to be easy? Build and Maintain an effective C&D strategy.
Steps to Implementation (per mckinsey):
Overcome the Barriers
Technical Barriers - lack of data or the ability to manage it effectively and accurately.
Organizational - lack ability to manage complex shift patterns, work hours agreements with staff, etc.
Get more for Less
Be Scrappy. You don't need to spend a ton of money building the perfect data system with business analysts / software. Many large scale companies and banks use spreadsheet tools and data drawn from existing resource planning systems. Can build an accurate data system in less than a month.
Eliminate Bias in Workforce Planning. Sales teams have a natural tendency to be over-optimistic in demand forecasts, leading to overstaffing.
Employee Happiness. Employees actually prefer flexible shift systems. Can be used as a reward for senior or better performing employees. Employees tend to be happier because they have work to do, but not too much due to a balanced staffing schedule.
2 | Determining SLA and Measuring Demand
2 | Determining SLA and Measuring Demand
Determining the Best SLA
Service Level need is dependent on the work type and customer
expectations. It is an agreement between the business / department and the receiving customer / department.
Service Level should aim for high expectations by determining whatis the market best and world class.
For call centers, the Service level is seconds to minutes, for
construction it can be days to months.
We will measure SLA demand on 2 different levels:
Same day or less SLA
Next day or Greater SLA
TOTAL HEADCOUNT NEEDED; NEXT DAY OR GREATER SLAs
Step 1: Build tools to measure and project weekly inflow. Know exactly what total headcount you will need at least 3 months in advance.
It takes days to weeks to months to get budget approval, open a req, interview, onboard, and train new hires. This time frame is much longer for some roles than others.
For companies with seasonality (aka the “solar coaster”), it is essential to plan ahead and move before the demand hits.
TOTAL HEADCOUNT NEEDED; NEXT DAY OR GREATER SLAs
Step 1: Build tools to measure and project weekly inflow. (cont.):
Get Sales Projections at least quarterly for the year. If you don't have sales projections, then determine them by average sales times sales agents onboarding / offboarding. Get the data for when they bring on or off-board sales team members and use this to project sales.
Determine close rate of sales projections (how much are they above below). Use this as the “Driver”. You can use other milestones in a pipeline / operation as the driver, and the closer your driver is to a worktype, the more likely it is to be accurate.
Record inflow of each work type as a ratio based on the “Driver”.
Determine the average amount of hours to complete for each work type.
Total these hours for the number of hours needed to fill by a workforce.
TOTAL HEADCOUNT NEEDED; NEXT DAY OR GREATER SLAs
Step 2: Use the Data to build an onboarding process that works.
Run the C&D data at least weekly, checking in sooner if needed.
Check for accuracy in the data by comparing with other sources.
Determining the onboarding and training period for each work type.
Each week look at estimated headcount need and hire in advance to accommodate this period. If the employee will only be 50% effective a few days or weeks in, account for this in this time period.
Develop overflows. This step is one of the most important. Sales and employees can stray outside of what you expect. They will do this by taking time off or having a shift in the market, economics of the business, sales model, product, etc. Much of the time there can and will be missed communication from either of these parties.
To solve this problem, develop “back-up” resources that can jump in and prioritize this work type. This can be members of another team, an outsourced
group that is piece-rate, etc.
Running incentives, working overtime, or throwing management in to a queue can also work, but will end up diminishing opportunities for process and
team development, both are an ROI for long term cost / sla improvement. If your department metrics are not improving, and your management team
spends the majority of their time in the queue, then the root cause can be ineffective C&D management.
SAME DAY OR LESS SLAs
Step 1. Build a tool to measure hourly inflow and a create a resource planning system (Staffing Schedule).
Determine when work is coming in and total in to a graph as an
average hourly inflow (get a 2-3 week period) of data. Make this rolling
if possible as it will adjust with seasonality, otherwise make it a
percentage of daily inflow. If work fluctuates depending on day of the
week, then break down each day average.
Total the hours staffed for when employees are scheduled to work, if
the schedule is too loose then a stricter working schedule will need to
be enforced to hit hourly or same day SLAs. Where SLAs can be
longer than same day, a stricter schedule is not necessary. Graph
these total hours along the same graph.
Look for gaps in overstaffing and understaffing. The gaps may
fluctuate on different days of the week. Overstaffing creates
unnecessary higher costs. Lower staffing creates SLA debt and poor
customer experience. Need to avoid both.
SAME DAY OR LESS SLAs
Step 2: Build a tool to measure hourly inflow and a create a resource planning system (Staffing Schedule) (cont):
Adjust Schedules through resource planning. Outline on the schedule. Negotiate with employees, make arrangements / agreements, develop reward systems for performance and seniority to pick a preferred schedule. Some folks may work longer or short days as needed.
Continuously Evaluate results. Further improvements in data collection and attrition / onboarding will need constant supervision. Will need to be able to project and accommodate quickly.
For unexpected C&D, build and implementing possible back-up options. The only alternative is having mgmt jump in the queue or acquiring a backlog. If management is in the queue it will like result in a lack of people or process development. This causes management dept for the department. Keep in mind that some teams only require a few hours of management support staff a week depending on the size / needs of the team, so acceptable to also be deep in the work most of the week.
4 | Continuous Improvement through Root Cause Problem Solving
Implementing a strong C&D system will still carry unforeseen issues. The leads to promises and agreements missed with Service Level.
Common issues could be: Market shifts for sales or employees, business acquisitions, filling in for mass attrition debt / inflow with other
departments, etc.
Root cause problem solve issues that are causing missed Service Level:
Identify if specific SLA for specific worktypes is being met.
If not, pull a sample of accounts (10-25) and deep dive in to them to find the reason they were not done in time, use the “five-why” analysis to determine what is the root cause of the issue.
Begin to Develop solutions that attack the root cause. Loop in any stakeholders / decision makers if necessary.
Agree with stakeholders on the best solution(s) to implement. There may be more than one. Determine action items for specific individuals, and set deadlines.
After implementation, evaluate the effectiveness. Is it solving the problem entirely? Is it not effective? (should be stopped as it is now a waste of resources). Does it need to be refined? Is it worth more than it costs to deliver?
If it is not effective, bring back to step 2 and repeat, if effective, keep that operation process / strategy in place.
Page updated
Report abuse