Cost it well and cost it right

IT services, especially maintenance and support, has always been driven majorly by cost considerations, besides, of course, the desired quality of service to support the business. The latter is an important aspect to take into account as IT became the "enabler of business rather than a support function" (... it's amazing to recall that this was even a debate a couple of decades ago and there were case studies written to establish the case for IT and its importance).

Price is just an extension of cost - it is predicting the cost that's the challenge. I was part of the global costing cell of Suzlon, where we were responsible for product costing and project costing. Building up the product cost involved adding up the following:

1. Gathering the cost of each component in the bill of materials (BoM) with inputs from SCM team's negotiated prices from vendors
2. Cost of manufacturing including labor and equipment
3. Logistics costs to ship the product to where it needs to billed at, the so-called Incoterms.
4. Overheads of all sorts

Projects could be (i) bounded - e.g., installation and commissioning of turbines, or (ii) ongoing - e.g., maintenance of turbines.

Similarly, IT too involves products (software, hardware / infrastructure) and services - sort of equivalent of projects - which could be various types - (i) building an IT system - predominantly making software on bought-out / paid-for infrastructure, (ii) configuring an IT system for use - involves setting up the whole software and infrastructure systems to work as required in conjunction with rest of IT, and (iii) maintaining IT systems

Costing for IT is a bit more complicated than in manufacturing, as everything in IT has tremendous amount of human effort involved and the various ways in which the effort can be delivered.

In case of products, the effort is incurred once and any number of copies of the product can be sold, which makes it less un-predictable as the projected price can be adjusted once the product is ready and the making-cost is incurred.

The real challenge is in costing for services as these predominantly involve human effort whose extent is difficult to estimate upfront. There are popular case-studies of failed ERP implementations which had massive cost over-runs (e.g., Nike). When companies outsource these by entering into contracts with IT service providers, the pricing challenge is passed on to the vendors. While these vendors are supposed to be well-versed with industry best practices and have expertise from multiple successes in the past, the inside stories are full of large-scale cluelessness and seemingly intelligent guesswork within range good enough to win business.

The company that wants to outsource - although it has better information - has similar challenge in fixing up its own budget. However, because of the closer view it has, it's expected to have better idea of what it would take to execute and the price it translates to, although many uncertainties still persist in the minds of its experts as well.

A vendor, on the other hand, consumes all the information that's made available to them and estimates the effort, costs the effort based on their cost base, plays with margins and gets to the price. But that's an over-simplification. Between effort and price, there are numerous aspects and related parameters that determine cost of execution of the project and the subsequent price, and it's not just about people and their effort but everything that goes into delivering the project. Here's an indicative list, which by no way is exhaustive.

As can be easily seen, there are a large number of variables one can play with. It's not necessarily wise to try and be deterministic about each of these, but to be able to decide at the solution stage would allow a vendor to deliver with a plan based on a budget. And to be able to model with such a wide array of variables and underlying options needs (a) a solutioning mechanism that allows architects to think, visualize and decide their solution parameters, and (b) a costing mechanism - a tool - that understands these variables, has an ability to discern and validate, and has the sensitivity and robustness to provide cost implications of each choice made.

I've had good fortune of working with many IT outsourcing service providers. Among those I've worked with, Accenture has deeply impressed me with the robustness in its solutioning processes, costing databases and tools, along with a rigorous review mechanism. It helps Accenture in many ways - (i) it ensures that many more aspects are given fair and careful consideration at solution stages, than just the effort estimation, (ii) it ensures deeper and holistic review where every choice parameter can be questioned, and is hence chosen with care, (iii) ensures greater optimization at many levels, (iv) ensures greater consideration to project deliverability and optimization to ensure the same, (v) helps identify, assess and plan for risks, (vi) helps prevent manipulation of solution parameters to get to lower cost, and (vii) helps drive high level of confidence in being able to execute with success within budgeted cost.

There is of course a measure of rigidity that creeps in when one works with such an elaborate process and tool driven model that everyone strictly adheres to. Companies must also balance this with flexibility to improvise, course-correct or make rational choices when the older assumptions turn out bad.

I must point out that the Indian IT service providers can do a lot better in their solutioning processes, costing models, tools and underlying databases for costing / pricing, and review processes. Greater detail to solutioning followed by better costing mechanism is essential for confident and successful delivery later on. It's unfortunate that most large Indian IT system integrators can be rated very low on these parameters. While it's important to win so as to be able to execute, not being careful with solution and cost will only render the win meaningless and as equivalent to trying to deliver later to a predetermined price that would be too tight, being generally low, and would require fewer people working longer hours, under stress and facing failure.

Not all unknowns can be figured out or accounted for at the outset, but an honest and detailed solutioning process, a robust costing model powered by efficient tools and a thorough review mechanism can together help vendors enter engagements with greater confidence and probability of successful delivery.

To conclude, just a thought - as AI agents replace human effort, perhaps we'll be able to achieve greater precision in spend on IT projects vis-a-vis budgets, in spite of uncertainties as they emerge. Would love to hear your thoughts...