How can you use historical WFM data to improve future scheduling?

Historical data of at least 2 years is ideal to provide seasonal or cyclical volume patterns for Schedulers to apply to provide adequate projections. Outliers in the form of pandemics or economic recession may influence data projections. However from agent productivity metrics the latest trending applies versus historical trend. This is attributed to current situations which could be cause by various factors work or non work related.

Have the historical trends upto 6months. This will give you a detailed view of your process performance. Have the hourly,daily trends of volumes inflow vs handled. Plot to understand the reasons for miss if any and do the course correction accordingly. Validating the trends frequently will ensure you smoothen the graph and also manage the costs efficiently. Agent utilisation and schedule adherence are the other two critical factors which needs to be validated and improved after a period of time.

Also review agent adherence. How well are agents following their schedule? This is the fastest and easiest fix if KPI's are being missed such as service level, speed of answer, and abandonment rates. Review if the interaction volume is truly off track or if the agents are not following their schedules which is negatively impacting the contact center KPI's.

One of the best approaches to identify patterns, trends, and anomalies in your data, such as seasonal fluctuations in a year is to gather historical data for not less than two years.

Another way of collecting the data is directly from the source of the hiring. Companies start hiring entirely based on the need of the managers and the projects milestone completion. If the work is repetitive, & the hiring trends in the company remains consistent YOY, then the quickest way would be to understand the timeline required to fill the same position. The data collected from them would give a broader view on both the quantitative and qualitative aspects. The agents/third parties engage only once they see a need, otherwise trends depend entirely on the source of hiring. Also, as per my experience, trends are dynamic at the source mostly. Once the data is collected, the scheduling/hiring time would be less resulting in costs as well.

The accuracy of historical data is best when the data range is small. This is because it is more difficult to identify patterns in data that spans a large range of time. For example, it may be difficult to see peaks in call volume that occur over a 1-hour range, but these peaks may be clearly visible in a 15-minute range. The disparity of calls over a given range also affects the accuracy of forecasts. If there are large spikes or dips in call volume, this can make it difficult to predict future call volume. For this reason, it is important to use data with a small range when making forecasts.

Most of the data is already available in the system as VMS/inhouse hiring systems are updated regularly. The way we would showcase / utilize the data in the time bound manner is something that we must focus on. Analyzing the data is most important than the presentation of it as the key lies in knowing the previous trends, accuracy of the data must rechecked before we start analyzing further needs based on the same.

Seasonality is key, adapting to moving holidays, making notes of ‘events’ whether sporting, weather, internal impacts etc-try and map as many as possible for modelling forecasts

Breakdown analysis into smaller chunks. I typically differentiate between 2 types of analysis: 1. plan for BAU, that is making sure we are adequately resourced based on the typical demands (therefore looking at typical shrinkage, handling times, etc). This is my biggest focus and helps me plan for longer term changes. 2. Understand how you have previously dealt with previous spikes. The effort here should be focussed around how well contingency plans have worked in the past and aimed at improving these.

I am not the type to use Excel to analyze things. Sue me. You can analyze and visualize data in WFM tool. USE YOUR TOOL FOR WFM, not for generating reports that are going to help other sectors in the company. Your WFM tool should help you create the story - why something happened and how to deal with. And once again observation is your friend :)

Ranking the goals according to their importance, urgency, and feasibility sums up this part of the process. We must also note that the all the parties involved in the WFM process are also aware of the goals as they play a major role in fulfilling the objectives; their inputs must also be taken into consideration before setting up the goals as they have been part of the process historically and have a complete view on the success / failure of the previously implemented strategies.

One of the popular goals is to reduce wage funds by 10% in the next quarter. Moreover, some goals can conflict for example to increase revenue vs. decrease wage funds. However, in some cases, it may not be achievable, and in order to increase revenue you need to increase wage funds.

Identifying and prioritizing WFM goals... Topic that I think is the biggest pain point. I do not define increase/decrease revenue and or changes in handle time or turnover as WFM goals. These are company wide goals impacting a lot of sectors not only wfm. For WFM it would be nice to prioritize improvement of Forecasting and lowering the Mean % Error, anticipation of things that are not impacted by incoming data - like shrinkages and similar things, and then scheduling. These can be goals that could be presented on long runs for WFM. Now if you include the RTA part, you could add more interesting short term things like response to changes, intraday efficiency etc.

This is key. Understand what changes you are planning and consider the impacts of these. You may be able to reduce handling times, but at what cost? You may be able to improve customer satisfaction, but at what cost?

A capacity planning tool can also be used. Look for a tool that can figure in your shrinkage, attrition assumptions, schedules, past historical volumes to save time in gathering the data to input manually and get more accurate results.

The workforce is a major contributor to production. Their skills and experience are a valuable asset to creating a viable plan to improve performance, service level, and overall productivity. Create a space for feedback from employees on the positives as well as areas that might require improvement. This can give you a different equally valuable perspective on what improvements would be most beneficial to the company's performance as a whole. Creating a similar space for supervisors/management to contribute and cross-referencing both would provide additional value.

Focus on a best case and worst case scenario. Evaluate the benefits and costs of those 2 scenarios. It is also valuable to have a good network you can rely on for discussion and challenge.

My favorite thing to do. Create scenarios and test them. What if analysis are the best and will prepare you for 90% of the things you will need to respond. The more you use this the better you will get. Scenarios will help you on the long run to improve working efficiency in the call center as you will have a thing to point to and say "here we anticipated this, so lets prepare better for the next time". Scenarios can include increase or decrease in volume, aht or fte's. They can also help you to show your organization what better staffing would provide as a result. Use it as a powerful tool. Use it for your weekly forecasts, and or for capacity planning. The least it will do is spark conversations.

To execute your WFM plan there is a convenient way to communicate with participants via mobile application. It includes information related to WFM goals, news, and plans

Implementing YOUR plan is harder then it sounds and looks. A lot of variables... One thing to note is - come prepared for everything, even the things you think are boring and will not be questioned - prepare for them. I don't like to use the term "selling" cause painting fits better. We are not sails people, we work with data, observations and facts. Prepare your plan, paint the picture and make sure that is simple and understandable as you would explain it to a kid and the kid would be able to reinterpret it and act on it. When it comes to communicating, make sure that you send an email to all the key stakeholders so for one you would have a paper trail, for two you would have it as a reminder for everyone.

Selling your plan will ensure it delivers as closely to the forecast as possible. Having buy-in across (both from the business at all levels and the support functions) is imperative.

With tools like MS Teams, sharepoint etc. it's really easy to make all relevant parties aware of the plan. Creation and implementation of the plan is a collaborative effort, so it is very important to involve all parties, especially those, that can execute the plan.

Have daily and weekly post-mortem reviews. Identify where things went right and where there were misses. Identify what caused the misses. Was it something controllable? Be able to adjust planned shrinkage accordingly and "Plan for the unplanned". We actually realized when we hit the targets, we were still 2% short of staffing in reality, so make sure you have a way to calibrate the WFM to adjust for Forecast vs Reality. We were able to add the 2% staffing increase by default going forward to make the WFM better for us.

Long term plan is never fixed and you need to be prepared to continuously review and adjust it based on the latest situation and projection. You need to be in regular connection with business leads to understand the current needs and shifts in demand

Even the best laid plans can go awry. Do not get discouraged if a plan does not work. Be prepared to change course due to new information, change in staffing, change in metrics, etc. Flexibility is the key to successful Workforce Management.

Historical data plays a crucial role in scheduling. It provides valuable insights into past patterns and trends, allowing organisations to effectively manage their workforce and determine the appropriate size of manpower needed. It helps organisations to identify peak periods of demand, seasonal fluctuations, and other patterns that impact staffing requirements. It helps in making decisions about when to adjust the workforce size, whether it's scaling up or down. For example, It can indicate when overtime should be scheduled if demand increase or considering temporary staffing for short-term situations. Utilising historical data in the scheduling process enables organisations to optimise resource allocation and improve efficiency.

Understanding the market you are planning in is so important. What are your competitors doing , will they influence your volume ? What are factors may influence, cost of living or other factors well out of your control . Also look carefully at your historical data and understand if there are anomalies that you need to factor out like out of cycle events . It’s really important to think laterally about your forecasts to help make them as robust as possible . Record all the factors you have thought about and continue to reassess . No matter how smart you are there will need to be changes .

Always consider seasonalities, holidays, events (product launch, start/stop of service from previous years). This will help in your forecasting patterns. Also the latest previous months, as this will show you latest trend. From these, you will be able to determine where possibly your future forecast will land on, and may be able to think of adjustments that may be needed to prepare for the "worst", if ever. +1 to all earlier contributors who mentioned that your forecast may not always work as planned but having been able to review all available data keeps you equipped and ready for whatever cimcumstance that will arise. Continue studying your histocals and trends and continue learning in improving your forecasts.