Last year like many people, I found myself laid off due to COVID-19. I suspected this would happen since my title barely reflects what I spend most of my time doing. Although I worked in a Call Centre, I hardly handled any calls. For those expecting me to have a lot of call stats, my numbers were probably disappointing. Arguably, I was a prime candidate for a layoff notice. Perhaps because people didn’t understand what I do, I soon found myself at home. The experience was not overly traumatizing. Having spent more than nine years at the same company, I was starting to feel like it was time to move on anyways.
There I sat the following week, checking my bank account as my first unemployment and stimulus benefits landed. Literally on that same day, I received a call to return. It was in the middle of the week in the afternoon. I was surprised to be expected back immediately. Even if I could get ready to leave on short notice, I quickly calculated that the commute back and forth would likely exceed the number of work hours available. I had stayed up late the previous evening checking out job postings. It was my duty to look for work. I was kind of curious if I could get a better job! I suggested that I could return on Monday. The problem from my standpoint was that – if they could change their mind that quickly (i.e. within days) – they could easily change it again. I wanted to make certain it wasn’t just a passing thought. I didn’t want an impulsive decision to adversely affect my eligibility for benefits.
Since my return, I have been sensitive to the fact that people might still be unaware of what exactly I do. I also noted how in the year before the pandemic, many people seemed convinced that I spent much of my time doing reports that could be easily automated. So I decided that I would make all of those automated reports useless – since I am largely not involved in their production anyways. Further, I would undermine their findings, making the guidance highly questionable. I would create new reports that are impossible to replicate unless the producer 1) can program proficiently; 2) has access the unpublished theory; 3) is not scared to push their workstation; and 4) is willing to speak authoritatively on something that requires a lot of explanation. Then I decided to exploit COVID-19 to specialize in remote metrics.
Percentage of Unique Callers
Consider a situation where 10 agents taking 10,000 calls in a month from a physical call centre are also found to take about 10,000 calls in a month working remotely. Are they doing the same amount of work? The answer is no – or at least, not exactly. Many people are familiar with the trade-off between quality and quantity: it is often possible to produce more if quality is allowed to slip. Without necessarily listening to every call, I should be able to quantify the level of call quality in a certain context – the level of functional or proper handling. In any event, this was my basic idea when I decided to write a program to determine the percentage of unique callers from the call logs.
For the sake of argument, let us assume that about 80 percent of callers are unique pre-pandemic and maybe 60 percent unique when workers are operating remotely. In other words, most of the time maybe about 20 percent of callers are repetitive pre-pandemic and 40 percent later. In order to understand how this might be possible, consider a scenario where more agents simply hang up on clients when they are working remotely. Seems a bit far-fetched? How about not fully assisting clients? When 40 rather than 20 percent of clients are calling back due to lack of adequate handling, this means that the agents are handling (10000 x 0.8) – (10000 x 0.6) = 2,000 or 25 percent fewer unique callers per month.
The percentage of unique callers can be examined in relation to agents, queues, or the call centre as a whole. In relation to agents, it can be interpreted as a performance metric. Because clients that are inadequately assisted seem likely to call back, the number of calls that clients make loses its value in terms providing business guidance. If 10,000 clients call in February with agents on-site – and 10,000 clients call in March with agents working from home – actually 25 percent fewer clients called (using the 80 and 60 percentages in my example).
Separating Business from Operations
There are stats, but there are also the underlying conditions giving rise to the stats. In order to resolve a problem, it is necessary to understand what brought it about. If a client found it necessary to call 10 times in a single day to deal with essentially the same problem, to me one option is to determine which agents seem to demonstrate the inability or unwillingness to deal with the client. Inability can be addressed with additional training. Unwillingness might be tackled through counselling and behavioural modification. Either way, it is possible to maintain a metrics regime to monitor this type of pathology and measure the effectiveness of remedial measures.
Although a metrics regime can be maintained, part of the reason I have taken a special interest in repetitive calling is that the regime is unlikely to be automated. It is necessary to manually examine each case and apply ontological criteria to generate data. My conceptualization of the problem therefore plays a large role in enabling the analytics. Indeed, most of the analytics that I now handle come from data that I collect myself. I have therefore become focused mostly on small data. One important task for me is to delegitimate big data that lacks my added value.
Rather than following the number of calls – a significant percentage of which is made up of repetitive calling – it is actually better to separate: 1) the number of unique calls (a business concern); 2) the number of repetitive calls (an operational concern). Previously, the business and operational challenges were conflated. This means that operational impairments and improvements could make an important aspect of business appear to increase or decrease. In fact, the more often clients are forced to call back, the more it seems business is increasing! Observers might point out that the company seems to be doing more work but making less money.
My basic theory behind social teratogenesis (in the context of this blog) is as follows: for the most part, clients are roughly about the same. Some might be identified as problem clients. Taken as a whole, they are not all that different. It would therefore be peculiar or statistically improbable for specific agents to be associated with more incidents of repetitive calling on a regular basis. The theory is that the teratology, rather than being intrinsic or internally conceived, extends from external conditions such as the interaction with particular agents.
My argument is that teratogenesis generally increases when agents are working remotely. Although it is true that a client could call 10 to 15 times even when agents are physically present in a call centre, I suggest that such cases are likely more common in relation to a remote workforce. This raises a number of deep issues about the importance of certain types of supervision and peer interaction – and the extent to which the erosion of normalcy can give rise to deviant behaviours.
At first glance, it might be tempting to associate teratogenesis with less quantity of product. But for the most part, I have been describing reductions in quality as an expression of teratogenesis. In an open environment, a reduction in quality might be interpreted in relation only to a specific client: e.g. John Smith called 15 times. The business might lose 1 client. In a closed environment, this impairment is more properly assessed in relation to the opportunity lost: e.g. John Smith called 15 times – therefore, 15 clients might be lost. By being forced to assist John Smith 15 times, not only might John Smith go away disgruntled, but 14 other clients were prevented from accessing support. So all 15 clients might leave disgruntled. Social teratogenesis is not an issue to take lightly.
Future of Remote Workforce Metrics
I feel that more people will be permanently working from home in the future. It is important not to assume that the standard metrics regime will remain effective when dealing with remote workers. Using the percentage of unique callers both to evaluate agent performance and better understand the business climate can help to improve guidance. Studying the pathology of repetitive calling will provide actionable insights to optimize the performance of agents.