Multi-skilling provides a great opportunity for efficiency savings, customer service improvements and adding variety and ownership for advisors. Multi-skilling also adds layers of complexity which can be very difficult to understand and accurate plan for. As we go searching for cost and efficiency savings we will often turn to multi-skilling and blending channels, but does anyone really understand the complexity of multi-skilling and the skills matrix?
There is no doubt that there are efficiency savings and potential customer experience savings by front line advisors being multi-skilled and in some instances across multi-channel. The ability to handle a range of enquiries with the same level of expertise and quality. This is especially advantageous to low volume skill groups as this supports the economy of scale. However, it is possible to add to many layers of complexity which result in configuration that no one (and I mean no one) understands. Experience will get you so far, but the reality is that you will struggle to accurately model a future scenario and truly understand the on-the-day performance.
The worst examples I’ve seen are where no effective skills matrix is in place, lack of understanding of workload delivery patterns and shift patterns not taking in to consideration the different levels of competency and mix of skills. Added to this, on-the-day the fire-fighting approach means that skill set changes are made to advisors based on “gut feel” and current interval queues and service levels.
Firstly, let’s consider the skills themselves. How much consistency is there in the process for each skill? If there is a set process, with standard timings, then you have a good chance of managing multi-skills. However, where there is variability in the processes and volatility within the handle times you start to muddy the waters. It is not uncommon for there to be a range of handle times, as advisors add their own “character”, when this is then multiplied across a range of different skills it cam make for a very uncertain operation.
Then there’s the random arrival patterns of calls. How predictable are call and workload volumes? If the workload arrives on a conveyer belt, with even distributed calls one after the other, then again you have a chance of managing multi-skills. However, if calls are driven by a mixture of external factors (e.g. weather, market conditions, cyclical) and internal factors (e.g. marketing, renewal cycle), you are likely to experience a mixture of quiet and busy periods, sometimes with calls queuing and other times with significant gaps between each arrival.
Skill Level Management
Next, we need to consider the ongoing training of our frontline advisors and ongoing process changes. How do we track performance and importantly, how do we maintain the level of competency for each person? How many skills can one-person handle? Is this the same for everyone? Who manages the skill set up and more importantly who has control over skill changes?
When skills are changed, how are these modelled and reflected in future projections? How are all skills considered when one skill is changed? In the worst examples, I’ve seen Real-Time teams and team leaders making ad-hoc skill changes based on their “gut feel” at that time. Typically, this will be based on the calls in queue, or service level during that interval and without any “what-if” scenario modelling for the impact of that change. Sometimes the change will be right, sometimes it will be wrong, but there will be no review to understand if it was the right thing to do, so ultimately no learning. If the change was felt to be right then the person who made the change will think they did the right thing, if the change didn’t work it will be someone else’s fault! Changing skills during the day, without any “what-if” modelling is essentially opening a “can of worms”.
Next, we have the shift patterns. If everyone is skilled to the same level of competency, with set processes, a predictable arrival pattern and a robust skills management matrix, then multi-skilling is for you. However, if there is a range in the level of competency how is this built into the shift pattern? How is this considered during a shift swap? How is the level of competency built into your WFM? Fortunately, WFMs can really support this and help us to model our requirement based on the assumptions we input. But, how good, and I mean accurate, are our assumptions?
Too often I see shift patterns produced without any, or very little, consideration for the mix of skills or mix of competency. Rules have been created to form rotational patterns, or fixed patterns and these determine the shift patterns. Then to add further complexity, staff are allowed to shift swap.
Finally, there is the shrinkage factor, and all the considerations for managing people. If you have a range of people with a range of skills your shifts need to consider this, or you will experience a range of performance depending on who is in at that time.
Having had experience of managing small skill sets I first noticed this on quiet late shifts, e.g. Friday night between 8 and 10. Consistently there were 2 people scheduled yet performance would range massively even with the same call volumes. Some individuals had quicker handle times than others. The difficulty was that with only 2 people covering this shift, working a rotating pattern which allowed shift swaps, considering annual leave, it was difficult to pair-up the best combinations.
In addition, more skills can require more coaching and communication sessions, especially when process are changing regularly. This can lead to additional shrinkage in your capacity/budget plans as well as a greater degree of skill required for your Team Leaders/Managers.
The Multi-skill Challenge
This article is not designed to scare anyone, but to highlight the potential complexity multi-skilling can bring. There are great examples of multi-skilling working, there are too many examples where multi-skilling is being operated without the right level of governance and understanding. With so much variability, it is very possible for each assumption to be wrong but overall performance ok; essentially cancelling each other out. As a simple test, look back at the last time you achieve service (not slightly over or under, but achieved) and look at individual actual results versus assumptions. This will provide a good “sense-check”.
- Document the processes for each skill with standard timings. Don’t just use the current average handle time, obviously consider this, but understand the outliers and understand what the target handle time should be to give the right level of service and quality. Just using the average is lazy analysis, the assumption should be the appropriate handle time balancing customer (expected service), colleague (realistic time) and company (cost versus service) needs.
- Educate the business on the volatility and variability of workload arrival patterns. If in an environment of changing delivery profiles explain this. It is critical that the business understands what is predictable and what is uncertain and how this impacts the day, the week, the month, etc. If you service level targets are based on an end of week, or month frequency this doesn’t mean you need achieve service level every interval, or even day. There will be calls in the queue at some points and there will be quiet times.
- Maintain a skills matrix with levels of competency for each advisor, regularly reviewing this with the individual, team leader and wider operation. If you have a multi-skilling environment this should be one of your key reports, if this is wrong, then your WFM will be wrong along with your future plans. As a “rule-of-thumb” if your skills matrix looks complicated, then it is. It should be easy to explain, easy to update and easy to use providing an accurate (not exact, but trusted) view of the future, along with development and improvement opportunities.
- Review shifts patterns and skills coverage at different time horizons. Consider the long-term view, understanding the impact of attrition and recruitment and time to competency for each skill. Review when shift patterns are locked-down (e.g. 6 weeks out), review short-term (e.g. 1 week out) and review and reforecast on-the-day. Most importantly, have some level of control over skill changes, with changes tracked and reviewed.
- Understand the impact of multi-skilling on your colleagues, especially in respect of shift-swaps and annual leave. Don’t limit your prized highly skilled multi-skilled advisors by limiting their shift swapping ability with only like-for-like advisors, but do consider the impact of changes and understand that not everyone is the same therefore different mixes of people with bring different results.
This isn’t easy to do, but hugely rewarding when achieved. There will be mistakes along the way, and just as you think you are getting somewhere there will be a process change, absence, attrition and some new recruits thrown into the equation. Create a set of guiding principles, which the wider business understands, then create some processes to measure skills at an individual level. Create simple “what-if” models that are easy to explain to key stakeholders, to build their understanding and trust. Understand which skills can be mixed and which ones require specialism.
If you have any comments/thoughts please share on the Linkedin article
Author: Phil Anderson