Measuring Value
A measure of success
Having for decades charged a fixed price for work done, IT services firms now need to start selling value. But how to measure it?
By most of the metrics that the IT services and business process outsourcing (BPO) industry measures up to, one of the sector’s largest global firms, Infosys, reported highly respectable results for the second quarter of this year. On October 23, the company, headquartered in Bengaluru but with offices in nearly 60 countries across the world, reported that AI work was now accounting for 10.5 per cent of revenue, up 29 per cent on the previous quarter. It added that it had also signed 5.2 billion US dollars of what it calls large deals (pointing out that this represented its second-best quarter for such things in five years) and that it had put a further 4,480 people on its payroll, something that parts the Indian trade press felt optimistic enough to declare a sign that IT jobs, fresh from being reported doomed, were in fact making a comeback. One of the two large deals Infosys reported (which its own people dubbed “a mega deal”) was worth one billion US dollars, with an existing financial services customer. Nevertheless, amidst the bombast the same set of results revealed that Infosys had in fact lowered its forecast for revenue growth this financial year to between one and two per cent. In April, it had been hoping for 1.5 to 3.5 per cent.
This gap between a quarter that appeared perfectly healthy on most of the standard metrics and a forecast that showed slower revenue growth warrants special attention, not least as Infosys is certainly not an outlier. Just about every IT services and outsourcing business runs on similar metrics, be it in India, Poland, Tanzania, or Uzbekistan (indeed, anywhere that the old Indian model has been replicated—and it has been replicated an awful lot). That’s because for the best part of 30 years, perhaps even more, the metrics of the IT services sector have been incredibly simple. A contract is won, people are hired to deliver it, firms are billed for their time, and in turn watch the numbers that tell them whether or not those people are busy. Jayesh Sanghrajka, Infosys’s chief financial officer, put it about as plainly as anyone could in April, when he was asked why the firm’s headcount had (at that time) dipped. Headcount, he responded, is a function of utilisation and of the volumes the company sees.
Utilisation is, in layman’s terms, the share of staff at a firm actually billing a client at any given moment. Other metrics might include revenue per employee, the size of the bench (a pool of employees who are currently on the company's payroll but are not assigned to any active, revenue-generating client project), or the shape of the company’s pyramid (usually with a wide base of graduates). These are measurements that the BPO industry has used for decades, and all, one way or another, measure volume, such as how much work is actually being done and by how many people. None of them truly measures what that work might actually be worth to the client who is asked to pay for it. If there’s a reason for that, it’s that it never particularly mattered as long as the amount billed for an hour of a developer’s time increased (if only slightly) every 12 months.
The metrics that the IT services sector has used for decades are sending out strange signals.
No wonder, then, that these metrics have of late been dispatching some strange signals. The Indian BPO industry’s trade body, National Association of Software and Services Companies (Nasscom) said in February that India’s technology sector grew its revenue by 6.1 per cent in the financial year that ended in March, to 315 billion US dollars. At the same time, it added only about 135,000 people to an existing workforce of nearly six million, representing growth in headcount of 2.3 per cent, well below revenue growth. There is a clear pattern of de-linkage between revenue growth and headcount growth, which (as well as anything) marks a transition from a linear, headcount-driven model to a non-linear, productivity-driven model. Customers, in other words, do not come to create jobs, they come to get work done.
Or rather, they come to get work done for a bit less than they used to pay. Staying in India, K. Krithivasan, the CEO of Tata Consulting Services (TCS; India’s largest IT firm, and Infosys’s biggest rival), told Reuters in August that about 80 per cent of the contracts in his firm’s finance, HR, and business-services arm are now priced on outcomes rather than hours. He went on to say that the productivity gains handed back to clients usually amount to 10 to 15 per cent over the life of a project. Sandeep Kalra, who runs Persistent Systems, a smaller and (lately) rather nimbler competitor, went further, saying that his clients now wanted the same work for 25 to 30 per cent less, “while expecting faster delivery and higher productivity”.
It should be possible to run a renewal through that same lens and come up with a few conclusions, primarily that a client pays (let’s say 25 per cent) less for much the same scope of work while the service provider, almost certainly tooled up with an army of with AI agents, does that work with fewer people and moves the rest on to another project (or, if there isn’t another project, out of the door). Revenue per head goes up as utilisation tightens, and nobody is left sitting on the bench. The margin, if the service provider has been even remotely smart, holds steady or even increases slightly. But the headline indicators (overall revenue, headcount) will not necessarily improve. Far from it, in fact.
Ready or not
Whether the industry is ready for a world in which the value of each unit of work matters more than the volume of it is a seriously moot point but in many ways the most crucial question facing the sector. The signals that once made it clear if there were choppy seas ahead are now being made irrelevant, and there’s precious little evidence to suggest that the sector can actually tell if it’s sinking or swimming.
Unsurprisingly, pricing is the most visible and problematic factor in the equation. A survey by HFS Research, an analyst firm which looks at the IT services and BPO sectors in more detail than most others, found last year that 42 per cent of outsourcing engagements were still priced on inputs (in other words, the old fashioned way: people and hours). Only one in five were priced on outcomes, although buyers did, to be fair, expect that share to roughly double within three years. Endava, a British firm with much of its delivery capacity in the Balkans and Latin America, told investors a year ago that outcome-based contracts made up 24 per cent of its revenue, with the other three-quarters still billed on time and materials. Globant, an Argentine-founded rival, has gone further than most by selling what it calls AI Pods on a subscription that it prices much like ChatGPT or Claude, by token consumption rather than by the hour. That is all well and good, but it still earns only around two per cent of its revenue that way.
As visible as pricing is, there’s also the question of the back office to contend with, almost all of which (timesheets, professional services software, resourcing tools) was built to record hours worked and people employed (and not without reason given that for 30 years that was what the services sector sold). SPI Research, which has been benchmarking professional-services firms for nearly two decades, reported this year that average billable utilisation across the 500 or so firms it surveys fell to 66.4 per cent in 2025, the lowest in its 19 years of records and proof that more work is being done in fewer hours, something that even the most sophisticated timesheet will not likely be able to pick up. (And nor, for that matter, will a billing rate, which measures what an hour costs, not how many hours a job now needs).
Problems always start before they show up in the numbers.
For so many firms (be they in the IT services sector or any other), problems always start before they show up in the numbers, which makes recognising institutional weakness before performance deteriorates so crucial for future resilience, and sometimes even survival. When even the headline numbers (which, eventually, albeit too often too late, might point at underlying issues) no longer make sense, the imperative to build an internal system of signals that can see what is actually happening becomes paramount.
How many companies actually have the necessary capacity for reinvention, to shift from selling what made sense yesterday (and may still make sense today, namely hours worked) to selling what tomorrow requires (value) remains to be seen. The first step is creating that internal system of signals that recognises where the real value lies, something that requires far more than yet another strategy, initiative, or even transformation programme (which are usually external facing). It will require wholesale changes in how work is valued, allocated, recorded, and billed within the company, because most of the current tools to measure these things fall short, having been designed for a world that is disappearing.
Ultimately, changing markets and changing realities requires a change in mindset, one that is prepared to pass on lucrative work if there is no long-term value in it. The IT services sector may be one of the first in which the penny drops, but it will not (and should not) be the last.
The byline Reinvantage Insight is used to denote articles to which several members of the Reinvantage insight and analysis team may have contributed.



