These days it seems every supplier’s infrastructure pitch book is full of the virtues and potential benefits of their drive toward automation, the objective being to get the same work done for less. What’s not clear is whether the supplier will actually be able to achieve what they promise or how to allocate the benefits between buyer and seller.
The same for less is a well-travelled road; the same goal drove moving work to less expensive delivery locations over the last couple of decades. Along the way some algorithmic alchemy created an acceptable balance among costs, margins, prices and benefit to the buyer. While the arithmetic to ensure the benefits were reasonably distributed amongst buyers and sellers could be complex, the factors of production to drive economic verification models were pretty well known, or at least could be with a bit of research. Underlying it all was a basic assumption, that an FTE was an FTE, and many buyers used the number of proposed FTEs to validate a suppliers’ ability to actually perform the work.
Automation changes all that. Is an FTE still an FTE, or is an automation assisted FTE a 125% of an historical FTE or maybe it is 150%, or maybe even more? What if there is no FTE at all just some robotics doing what an FTE used to do? Since an automaton is likely to make fewer mistakes than a human FTE, and will do those error-reduced tasks faster than the human FTE, the promise of better and faster and cheaper seems attainable.
Nothing wrong with any of that — it all sounds pretty terrific…
Yet the road to automation nirvana features unexpected curves and potholes. Overcoming these obstacles requires answering the challenge of accurately projecting the financial impact of automation. Putting aside for a moment how to allocate the benefits of automation between the buyer and seller, what will a supplier expect costs do over the typical five-year term of an ITO? What is the probability of the supplier under and over estimating progress?
Suppliers projecting year-over-year pricing improvements is nothing new. ITOs have, for many years, included cyclical pricing improvements in the 4% to 10% range, linked to known learning curve improvements and the introduction of management tooling. But automation presents two new problems, the automation technology is just now being deployed into supplier delivery engines and processes and the potential range of productivity increases is far larger. Assuming, sans automation, a provider server administrator can oversee about 75 virtual server images, resulting in a gap of at least one and half orders of magnitude with the productivity ratios achieved by Internet scale operators like Google and Amazon. Drawing a pricing curve between those two productivity points is very different problem than computing the incremental adjustments suppliers have accommodated in their past pricing and will be, at the outset, far more difficult to model.
The inclination of buyers is to press suppliers to maximize improvements in year over year pricing over the term. Consider the consequences if a supplier makes an extremely aggressive pricing choice to win the business and later fails to be able to meet their automation goals – resulting in higher supplier costs. That leaves the supplier with several unpleasant choices, take a margin haircut, negotiate with the buyer for a price increase or reduce the manpower to the levels projected as if the automation objectives were achieved and run the risk of failing to meet service levels. Historical behavior would suggest a higher probability of the supplier reducing the amount of labor and taking the risk of reduced performance.
So how does a buyer cope with the situation of entering into a new or renewed ITO arrangement over the next two years or so before the automation track record is established? What kind of pricing improvements should be demanded? How does the buyer avoid unintended consequences and resultant issues in performance or price uncertainty?
Recently we have seen situations where suppliers, in order to maintain or expand margins, have reduced staffing across their entire delivery engine, resulting in individual buyers seeing resources simply vanish from their engagements, without suitable replacements in either numbers or skills.
Automation offers suppliers a sexy cover story for these moves. One can imagine hearing something along the lines of “of course we reduced staffing, that is the result of our automation efforts and that is how we were able to offer you the low pricing levels that you enjoy.” Skepticism would suggest that buyers would be hearing that refrain whether or not the supplier has actually achieved the automation objectives supporting that position.
The dilemma is that it is still early days and neither buyers nor sellers have the ability to accurately predict the automation benefits that will be achieved over a five-year term. What is needed is some sort of commercial mousetrap that can be adjusted over time; adjusted as more is known about the benefits of automation and the supplier’s ability to actually harvest them.
One way to build that adjustable mousetrap…
An appropriate approach would be to set several trigger points for discussion of adjustments, these triggers should include: (i) annual periodic reviews of the suppliers overall progress in respect to achieving the productivity changes to support the scheduled contract year pricing reductions, (ii) any supplier action to materially change the amount or quality of staffing on the engagement, and (iii) any sustained deterioration in service delivery. The first two of these of these triggered discussions should be conducted well in advance of any action to allow the buyer and supplier ample opportunity to resolve any differences of opinion prior to the supplier implementing their proposed or scheduled action.
These discussions should be formalized and should include buyer and seller leadership above the managers running the day-to-day relationship. Our experience is that improperly planned resource reductions are a cause for serious operational concern and have the potential to be escalated to the COO/CEO even in very large enterprises.