As if the tension from the previous post wasn’t enough, in this instalment we will ride the roller coaster of demand management. While the CM equation is pretty simple, given the inherent variation on both sides, coming up with the numbers is anything but. True to the maxim “variation is the devil incarnate”, trying to smooth out the peaks and troughs will make life a lot easier. So in this post, I’ll share some thoughts on the demand side of the equation.
To Forecast Or Not To Forecast, That Is The Question
As I mentioned in the previous post, there is a school of thought that suggests there’s no point in forecasting demand, as your chances of getting it right are pretty minimal. Instead, all of your energy should be focussed on building flexible and responsive capacity.
I disagree with this point of view because the act of forecasting and reviewing helps you gain a deeper understanding of your customers and your processes. So with this in mind, I jotted down some notes on forecasting. 3 pages later and I’d barely scratched the surface, so I will make demand forecasting the subject of a separate post. Suffice to say, if you don’t have a forecast you really are flying blind.
Demand Profiles – And You Shall Be Rewarded
Some may struggle with this idea, but management accountants are truly blessed. One of their major tasks is month-end reporting, where there is a spike in workload around the end of the month and the first few days of the following month.
The good news for practitioners is that not only do they know it’s coming, they know when it’s coming – it happens every month and funnily enough at the same time every month. They may not be able to predict when the CFO will decide to change the format of the reports or a restructure that requires a re-alignment of cost centres, but these are relatively rare events. Hence the demand pattern is well known, which makes the task of managing capacity relatively straightforward.
At the other end of the scale, consider those teams involved in settling currency trades. Trying to forecast their movement and the impact on demand is practically impossible given how unpredictable financial markets are.
This highlights the scale of the task at hand – it’s driven by the relative mix of predictable and unpredictable variation in the workload. Those functions whose work is highly predictable, even if the swing from peak to trough is large, have a far easier job of planning capacity than those where there is a significant random element at work.
So a useful place to start is to visualise demand patterns of how work comes into the team. Assuming you have the relevant historical data, creating the charts is pretty easy. It’s important to look at the patterns over a range of time frames – intraday, day/week, week/month or day/month (depends on the operating rhythm of the business) and month/year – as each will present a different challenge, with different treatments available to help solve the problem.
The charts below all represent “expected work-in”. As you can see, all of the charts display more than a modicum of variability, but if you have a known pattern then you have a base from which to work. improved.
A word of caution, be careful to read the scale – the week/month chart makes it look like weeks 4 and 5 are significantly quieter, where the volumes are only 3% lower – enough to make it worth acting on but certainly not alarming.
Once you’ve created charts like this for your teams, the scale of the problem becomes apparent. Compare this to the charts for the Month End Reporting team:
Whilst the week/month profile poses a challenge, once you’ve figured out how to manage this period, it’s relatively straightforward to apply this approach equally across the year given how smooth the month/year profile is.
In essence, the smoother the chart, the easier it is to manage. If you don’t have a smooth chart, then the fun begins. How do you adjust your capacity to meet this pattern? Do you hire enough people to meet the peak and have them check-out You Tube videos during the quiet periods? Do you only hire enough to meet the minimum and hope customers will forgive you? Do you hire and fire with alarming regularity? We will look at how to flex capacity in the next post, but the first task is to look at options for smoothing out demand.
NB For those functions where there is a significant amount of unpredictable variation, it still makes sense to strip out the one-off events to see the underlying demand profile. We’ll deal with the random spikes later.
Demand Management Options
Managing demand is challenging in a typical service environment, to say the least – it’s very difficult to try and persuade customers to send in their tax assessments before the end of the financial year – and it’s unlikely to be the only way you solve the problem. However, there are a few options:
- Pricing: this is the “happy hour” scenario and can be positioned as a carrot e.g. a discount to the base price or a stick e.g. a surcharge to the base price. In the Intraday scenario above, offering customers who send in their work in the morning a discount may shift the afternoon peak.
- Promotions: in the month/year chart, running promotions during the winter(southern hemisphere)/summer(northern hemisphere) months could help smooth out the trough, but it very much depends on the type of service you’re providing.
- Specifications/Service Level: you can change the service level to offer a fuller/faster turnaround time for customers sending their requests in Monday-Wednesday for the Day/Week chart. Intraday basis service levels, such as “order by noon and get same day service” are quite common, or offering a free wrapping service in a retail outlet during quieter periods. As a side-note, be extremely careful of setting “same day” service level expectations – essentially you are saying that you can reduce the time to manufacture as the day progresses. A customer sending in a request at the start of the day gives you say 8 hours to produce the service, whereas the customer sending it at the end of the day gives you minutes to manufacture the service. Destined to fail!
- Reservation Systems: booking systems obviously help smooth out demand but do present additional problems such as “no shows” and cancellations when it’s too late to re-book the same slot. These tend to be restricted in a service centre environment to specialist resources.
- Large Customer Negotiation: where demand is driven by a relatively small number of large customers, negotiating one-on-one to smooth the demand out is feasible and can make life far easier, albeit there may be constraints on the customer side.
- Complementary Services: the first five options obviously require input from product, marketing and sales functions, whereas this option can be effected directly by the Ops team. The trick is to find other teams that have complementary demand profiles and cross-skill. For example a team that was busy in the morning and then quietened down in the afternoon is a perfect complement to the intraday profile above. I used to manage an Inward and an Outward Payments team. The Inward team received a lot of their work overnight which was piled up when they arrived in the morning, while the Outward team received most of their work in the afternoon. Combining the two teams effectively smoothed out each team’s peaks and troughs.
All of these approaches can help smooth out the peaks and troughs, but in my experience it’s unlikely that they will remove all of the bumps. So if we can’t smooth out the peaks and troughs do we:
flatline capacity and accept that there will be times when there isn’t enough work and times when the team are overworked and service levels are impacted, or try and “chase” demand i.e. create a capacity profile that matches the demand profile?”
The first approach is certainly the easiest and in practice is the default position in most organisations I’ve worked in. The finance function essentially set both the FTE and the personnel expense budget, based on a notional top-down target that rarely has anything to do with demand. For me this represents both a lack of trust between the functions and a lack of confidence and maturity in the operations function to demonstrate that it can actively manage capacity. The root cause of this state of affairs is a disconnect between treatment of a request to increase and a request to decrease resources: “what CM giveth, finance taketh away” – in other words, the manager doing the right thing and releasing excess capacity soon learns a very harsh lesson when they try to add capacity to meet a demand spike and approvals are blocked.
In theory, if you choose option 1 you can also choose where to set capacity relative to the demand profile e.g. at the lower end or the higher end or somewhere in between. The choice will typically depend on how critical the service level is and how much it costs to bring on extra capacity. Most teams I’ve encountered have been in the upper quartile. Unfortunately, many get progressively ground down, year after year, when they are set targets they have neither the skills nor investment to achieve and resort to running the lawnmower over the top – as a colleague many years ago said – “every 3 or 4 years you just need to blow your plan so you are allowed to re-baseline”.
And on that bombshell, I’ll leave you until next week’s denouement (spoiler alert – its option 2).