In 2004 Chris Anderson’s now famous Wired article introduced the World to the Long Tail of niche marketing. This is a frequency distribution — a graph, that can be used to understand a retailing strategy for selling a large number of unique items in individually small volumes.
He showed how the internet, and the likes of Amazon, could support a near “infinite” shelf strategy, offering niche products on a vast scale. Each product line may sell in tiny quantities but added-up across the long tail, they become as valuable as high volume best-sellers. Trying to pick winners so as to stock a “finite” physical store seems pretty primitive in comparison.
In 2011, IT analysts at Forrester Research made an analogous observation for automation requirements within large service organizations — like banks, telcos, retailers, or utilities. Fascinatingly, the graph offers the same rich insight, but this time into automation and IT strategy. I will start by declaring that automation for businesses is good.
Looking at an organization through a “Long Tail” lens of tasks, there are a few high-volume repetitive tasks supported by an expanse of lower volume duties all the way down to single one-off decisions. Once laid out in this way, we can apply “automation fulfillment” to see what support the business gets. At the far left of the curve there are high-volume repeat tasks like payroll, accounting, finance, order management, and HR. Think of the myriad calculations involved here.
Back in the 20th century, payroll used to have vast armies of people fulfilling, calculating, making corrections, stuffing envelopes, adjusting for holidays, making further corrections, adjusting tax and of course correcting corrections. Nowadays, this is automated. In fact, these were some of the first areas to be automated and we’ll call them “Core IT”. Many times the systems in place are the originals and come from the eras of IBM, SAP and Oracle.
Moving along the graph we see lower volume “Strategic IT,” things like call centers, CRM systems, management information, analytics, BI, data publications, web support, process management and workflow. Interestingly, purchase decisions here tend to be dominated by compatibility with “Core IT” (the business may choose any system they like but their IT colleagues insist it fits in with their IBM or SAP choices of yesteryear).
Moving further along we enter a new region where task patterns become less frequent and the system resources become somewhat patchy. Often seasonal squeezes and operational runs leave this area exposed, requiring extensive manual work and temporary staffing to address bursts of activity. In other words, we are leaving IT country and moving into people territory, the real battleground between IT and operations. This is where the tension is highest between IT’s need to do things right and operation’s need to get things done. It is where life-long animosities are forged and where careers have ended
Once out of the tension zone we emerge in the domain of operations, or manual land. Here we have tasks that are resistant to automation (talking to customers) or so varied and heterogeneous that business cases can take hold. Its here we have call-center staff, fixers, data entry, invoice handling, doers, error correctors, regulation watchers, competitor response, and the pantheon of people doing things that make a business work. The character of this work is dominated by low-pay and high staff turnover. It’s the hunting ground for outsourcers (“your mess for less”) – things like F&A, invoice management, admin updates etc. Because customers are not controlled by “Core IT” they choose alien incompatible systems –some even use paper.
As a consequence, Forrester suggest 50 percent of all automation opportunities are being missed. And moreover this outer region of operations is the prime hunting ground. It is also clear that a command-and-control IT has not and, I’d suggest, will not, be successful in automating within this area. It is a situation akin to the old days of the Soviet Union where central planners tried to keep pace with the demand for fashion items like platform shoes and mini skirts. By the time the planners responded, the fad had moved on. IT departments are good at core but hopeless in these fast moving outlands.
Forrester’s recommendation is to democratize the work. Let the workers have their own “on-demand” automation capability using Software Robots.
As I have been arguing, Software Robots represent something disruptive. They mimic humans, but live in the cloud or on servers, they are the ethereal cousins to their mechanical counterparts. They are trained, not programmed, and carry out tasks which are rules-based and may involve interaction with multiple incompatible core systems (just like people do). Because robots use the same interfaces as humans they are “always” compatible — an unbelievable technology advantage that promises an explosion in opportunity.
The robot works in a team to carry out medium and large automation tasks. It supports automation, for example, where none was considered necessary — like handling errors. It is nimble, dynamic and versatile. It is the automation equivalent of the “long shelf” in niche retailing. It means workers become their own agents of automation and therefore fulfillment.
Peter Drucker, the management guru, raised the issue of white collar productivity a long while back, stating that manual workers through automation had achieved a 50x increase in productivity, whereas office jobs have remained static. Robots could change this.
Jason Kingdon is chairman of the robotic automation company Blue Prism.