The first half of the year closed yesterday as the worst six months for equities since 1970. Capital has been repriced, the growth-at-any-cost instruction that governed the last decade has been withdrawn in about eight weeks, and every board paper now arrives with an implied question attached: what does this actually return? Outsourcing is squarely in that beam. Business process outsourcing decisions have historically been made on a rate-card comparison, approved once, and then never revisited, because the organisation that would have done the comparison no longer exists. That is now changing, and the conversation has moved up two levels — from a procurement review to a board discussion about whether a large multi-year commitment is delivering what it promised.
The moment you outsource, you destroy the control group
This is the central measurement problem and almost nobody names it. On the day a process transitions, the in-house baseline stops being observable. The team disperses, the volumes change, the scope drifts, upstream systems are modified, and within eighteen months there is no defensible way to say what the alternative would have cost. Which means the ROI conversation has to be designed before transition or reconstructed painfully afterwards. The designed version is cheap: freeze a baseline artefact — the full cost stack including supervision and premises, twelve months of volumes by transaction type, the service levels actually achieved rather than the ones documented, the error and rework rates, and the exception categories — signed by finance, dated, and stored where the contract is stored. That document is the only thing that makes a later ROI review honest rather than rhetorical.
Four costs that never make it into the case
The retained organisation. Somebody has to manage the provider: governance meetings, service reviews, escalations, query resolution, invoice validation. This typically runs at a meaningful percentage of the outsourced cost and is almost always understated at the outset, because the roles are absorbed into existing jobs rather than created explicitly. Transition and knowledge transfer. Dual running, travel, documentation, the productivity dip on both sides, and the attrition among the people being transitioned out who leave before the handover is finished. Change requests. Anything outside the frozen scope is priced in a market with one bidder. The pattern is predictable: a tight price at award, then two years of additions at a rate nobody competed for. The coordination tax. Your own staff answering queries, correcting upstream data, and chasing items that used to be resolved by walking across the floor. This is genuine cost, it is never captured, and it is often the difference between a good case and a bad one.
The pricing model is now under stress from the other side
Most contracts running today were priced when wage inflation in delivery geographies was modest and attrition was normal. Neither is true this year. Providers are absorbing higher wages, sharper turnover and rising infrastructure costs against fixed unit prices, and the effects arrive as quiet substitution: a less experienced team on your account, slower resolution of the awkward cases, and a more aggressive stance on what counts as a change request. The naive response is to hold the provider to the letter of the contract and watch quality decay. The better one is to reopen the pricing mechanism deliberately — trade term length or volume commitment for an indexed price with a cap and a productivity offset, so that both sides know what happens next year rather than fighting about it.
| Cost named in the article | What to include |
|---|---|
| Retained organization | People and responsibilities the buyer still carries. |
| Transition | The work required to move and stabilize the service. |
| Change requests | Scope changes beyond the original agreement. |
| Coordination | The buyer's work managing handoffs and exceptions. |
Qualitative summary of this article's source text, not a measured outcome or performance estimate.
Four numbers a board can actually use
Total cost to serve per unit, all-in, counting the provider invoice plus the retained organisation plus the coordination tax. Quality, expressed as rework rate and the proportion of transactions touched more than once. Cycle time, end to end, from arrival to completion, not the sliver the contract measures. The cost of an error, because a cheaper process that produces more late payments, duplicate payments or credit notes is not cheaper. Report those four against the frozen baseline annually, with finance signing the arithmetic. Everything else in an outsourcing scorecard is activity data.
Three reasons to outsource, ranked by durability
Labour arbitrage decays. Wage convergence, currency movement and the cost of managing distance erode it every year, and a case resting on it alone gets weaker with time. Capacity elasticity is real and underrated. The ability to absorb a peak, a new entity or an acquisition without hiring is worth paying for, and it is the reason most successful arrangements survive their own business case. Capability you cannot hire is the most durable reason of all. If the provider does something you genuinely cannot build — a statutory specialism, twenty-four-hour coverage, a system you would never buy — price becomes secondary, and the relationship holds through a downturn. Be honest about which one you are actually buying, because the answer determines whether the current squeeze is a renegotiation or an exit.
Practical Guidance for BPO ROI Assessment
- Reconstruct or retrieve the frozen baseline before any review; without it, the discussion is opinion.
- Cost the retained organisation explicitly, by name and by percentage of effort.
- Add up three years of change requests and compare the total to the original savings claim.
- Measure cost to serve, rework, cycle time and error cost — and nothing else at board level.
- Reopen the price mechanism before quality degrades, trading term or volume for an indexed cap.
- Separate arbitrage, elasticity and capability in the justification, and test each one separately.
- Fix upstream defects that create volume before paying a provider per transaction to absorb them.
- Model exit properly — knowledge, systems, people and a realistic rebuild timeline — before threatening it.
The Regional Angle
Three regional factors are moving the economics of these contracts right now, and the first is pure currency. Gulf buyers contract in dollars or in pegged currencies, while delivery costs sit in rupees and pesos. The rupee has weakened materially this year, which means the real cost of an Indian delivery team has fallen in dollar terms for reasons entirely unrelated to productivity. If your contract is priced in dollars per transaction with no currency mechanism, that gain accrues wholly to the provider — which may be defensible, since they carry the risk when it moves the other way, but it should be a negotiated position rather than an accident. Ask for the currency basis behind the price, and if the exposure is one-sided, propose a sharing band: movements beyond a threshold adjust the unit price in both directions. Providers resist this less than you expect, because it also protects them. The second is tax, and it reaches the captives rather than the contracts. Federal corporate tax applies to financial years starting from the middle of next year, which turns every intercompany shared-service arrangement into a transfer-pricing position. The cost-plus margin that a group captive charges its operating companies has until now been an accounting convenience; from next year it is a taxable outcome that must be documented, benchmarked and defended, and the treatment of free-zone entities in these structures is still being clarified. Any comparison of captive versus outsourced economics made this year and excluding tax is already out of date. Get the transfer-pricing view into the model before the board sees it. The third splits one contract into two. Saudi Arabia is steadily requiring that work for Saudi entities be performed inside the Kingdom — through headquarters policy, procurement preference and nationalisation requirements applied to the provider's own workforce. A provider delivering your Saudi processes from Riyadh carries a fundamentally different cost base from the same provider delivering your Emirati processes from Manila, and no blended rate card describes both honestly. Groups with entities on both sides of that line should price and govern them as two arrangements, and should expect the Saudi-facing portion to offer much less arbitrage and much more strategic necessity.
The objection worth taking seriously
The strongest objection is that outsourcing ROI is theatre. The counterfactual cannot be observed, scope has drifted beyond recognition, and a board that demands a savings number will receive one that was reverse-engineered from the answer. Meanwhile, reopening a functioning contract in a downturn consumes exactly the management attention that should be pointed at revenue, and risks destabilising a process that is currently working. Better, on this view, to leave a competent provider alone and spend the quarter somewhere that moves the top line. The first half of that is correct. Once the control group is gone, any claim about what the alternative would have cost is a construction, and everyone in the room knows it. But the useful measurements do not require a counterfactual. Cost to serve per unit, rework rate, cycle time and error cost are all observable today, comparable year on year, and sufficient to tell you whether the arrangement is improving or decaying. And the purpose of the exercise in this particular market is not to litigate a decision made in 2019. It is to price the next three years correctly at a moment when the provider's own cost base has shifted underneath a contract neither side has reopened — which is a conversation that happens now, on your terms, or in eighteen months on theirs.
Common Questions
Should we bring work back in-house?
Rarely as a first move. Rebuild cost is high, the knowledge left with the people, and in-sourcing under budget pressure tends to recreate the original problem. Re-baseline scope and pricing first; consider in-sourcing only the judgement-heavy portion.
What is a realistic retained-organisation cost?
Higher than the business case said. Count the actual hours your managers spend on governance, queries and invoice validation for one month and annualise it — the result is usually uncomfortable and always useful.
How do we handle a provider asking for an increase?
Ask for the cost evidence, and use the request as the opening of a two-way negotiation: indexation with a cap, a productivity commitment, scope clarification, and a longer term if you want price stability.
What should we expect over the next twelve months?
Expect providers to approach clients with increases in the second half as wage pressure works through their books. Expect shorter contract terms as both sides become wary of long commitments in a volatile cost environment. Expect automation-funded pricing proposals, where the provider offers savings in exchange for keeping the productivity gain — read those carefully. Expect regional captives to be re-examined as tax structures rather than as operating models. And expect boards to start asking for the retained-organisation cost by name, which is the single question most outsourcing reviews are least prepared to answer.
BPO ROI Assessment — we rebuild the baseline that disappeared at transition, cost your retained organisation honestly, and put four board-legible numbers behind the next three years of the contract.