Follow the promise beyond the completed task

In 2020 I was operations director when a review of sixty-four individual sewing machines found eleven with two order claims each. The other fifty-three had one claim. That meant seventy-five claims against sixty-four physical units, not eleven orders in total. I had accepted a process in which local confirmations could appear valid without resolving their competition for the same machine.

The sales team wanted to preserve customer confidence and avoid an abrupt stop to promising conversations. I understood that concern, but allowing further confirmation before reconciling the assignments would make the operating contradiction larger. We held additional confirmations while customer-facing colleagues worked through the actual options. The case does not invent cancellations, compensation or a successful settlement for every order. It also does not describe a machine being physically dispatched twice.

We allocated ₹29 lakh to reservation redesign with the relevant commercial and technology owners. My responsibility was to define the operating commitment clearly enough that one team could not rely on a unit already promised elsewhere. An aggregate stock figure could not resolve which customer a particular physical machine could serve. The correction needed to connect the assignment and the promise, rather than ask colleagues to make a misleading summary balance through local adjustments.

The immediate outcome was a held confirmation process, reconciliation work and a funded redesign. I retain the original approval in the account because a local status had persuaded me before I understood its enterprise consequence. This is not a fraud finding or a description of a particular software algorithm. It is an operating lesson about the difference between a record that appears complete and a customer commitment that can actually be fulfilled with the physical resource available.

In 2022 the repair account showed 180 open jobs. A proposal to add bench capacity treated the whole total as technical work waiting to be done. When we examined the state of each job, seventy-two were awaiting customer authorisation of a quotation and 108 were authorised repair work. Those groups required different next actions, even though both remained open in the summary.

Technical colleagues were concerned that the aggregate made them appear responsible for delays they could not resolve by starting another repair. Customer-facing teams also needed a clearer account of which quotation required contact and what had actually been agreed. I asked that the work be described by its next permitted action. Joseph Mevqari retained technical assessment and repair-release authority; operations could not use a backlog discussion to replace his judgement or a customer's consent.

We allocated ₹34 lakh to workflow, status and customer-contact improvements. The programme was not permission to infer agreement from silence or to begin unauthorised work. It also did not prove a cycle-time reduction merely because the status account became more precise. We needed a usable distinction between waiting for a decision and carrying out approved work before judging what additional capacity, contact or process change might be appropriate.

The result recorded here is a corrected operating account and funded response, not an assertion that every open job was completed. I would not use the seventy-two to dismiss the remaining 108 as an insignificant queue. Each group still needed attention. The case shows why a COO should understand the meaning of an unfinished-work total before assigning blame or buying capacity: the same open status can conceal a customer's decision, a specialist task and very different responsibilities for moving the job forward.

A 2023 review followed sixty urgent spare-parts orders with a four-day customer promise. Forty-six missed that period. A separate fleet report showed 98% of scheduled trucks departing on time, which initially made the service problem look inconsistent with the transport account. The measures described different populations and started their clocks at different points. Parts could wait for consolidation before a truck entered the departure schedule.

Transport colleagues reasonably valued reliable departures and efficient loads. I did not ask them to abandon those measures. I asked that the operating decision include the waiting before dispatch became a scheduled activity. A customer did not gain a shorter wait because the eventual truck left punctually. The service promise needed a route appropriate to urgent parts, with commercial and technical owners retaining the relevant terms and approved-item requirements.

We funded an ₹85 lakh small-parcel service programme. In the following defined sixty-order cohort, eighteen orders missed four days, compared with forty-six in the earlier group: twenty-eight fewer late orders between those observations. The cohorts were not a matched causal study, and the result does not establish a universal service rate or prove that every difference came from the programme. It does show why the complete customer interval belonged in the review alongside fleet punctuality.

The case changed our operating discussion from defending one internal measure to understanding the sequence behind the promise. We retained the remaining eighteen late orders rather than describe the programme as a complete solution. The ₹85 lakh was programme funding, not an achieved saving. I would use both the observed improvement and the unresolved cases to guide the next review, while keeping scheduled-departure reliability distinct from order-level compliance with a customer commitment.

In 2024 a proposal sought ₹90 lakh for another demonstration room. Two existing benches appeared heavily booked, suggesting that more physical space would increase service capacity. A review of eighty paired bookings showed a different constraint: thirty-four pairs could not be delivered concurrently where the two benches shared one qualified demonstrator at the overlapping time. The missing resource was not established simply by counting available benches.

Store colleagues wanted a visible response to customer waiting and could readily explain the appeal of another room. Technical head Joseph Mevqari asked that readiness be understood as the combination of appropriate space, equipment and qualified people. I supported that distinction. A room could be completed while the same person still had to serve two concurrent bookings, leaving the customer promise unchanged despite a significant capital commitment.

We rejected the third-room proposal in that form and funded ₹22 lakh for appointment coordination and technical preparation. Qualified owners retained training and readiness decisions. The allocation did not create a second qualified demonstrator by itself, certify anyone or prove that all future conflicts had been resolved. It gave the operating and technical teams a more relevant programme to develop before treating additional property as the answer.

The outcome was a different resource decision based on the actual service combination. I do not claim that the benches were defective or that physical space can never constrain the business. This case showed a particular overlap that more space alone could not solve. A COO should be able to challenge the most visible proposed asset while respecting the specialist judgement about who can deliver the service. The useful capacity measure is the appointment that can be fulfilled, not the number of places in which it might occur.

A 2025 trial kept fourteen shops open from seven to nine in the evening. The specified period comparison attributed ₹18 lakh incremental sales and ₹6 lakh associated gross contribution to those hours. Additional staffing and utilities were ₹11 lakh. That left a ₹5 lakh gap for the defined trial period before other unchanged costs, rather than the attractive contribution to enterprise profit that the sales figure alone might suggest.

Commercial colleagues valued the additional access for customers unable to visit earlier. Operations also needed to account for the people and resources required to provide that service. I did not treat the ₹5 lakh gap as proof that every later opening was wrong, but I opposed automatic network rollout from the ₹18 lakh revenue headline. The comparison had its stated scope and assumptions; it was not a controlled causal estimate or the company's full profit account.

The CEO declined automatic rollout and approved ₹16 lakh for a selective follow-up review. We wanted location and customer needs to inform the next decision, with finance retaining the relevant contribution definitions and people colleagues involved in actual staffing arrangements. That funding was review expense, not a realised return or evidence that a later model had already become profitable. The trial still offered a customer-access benefit that needed to be considered alongside its cost.

My contribution was to keep both sides of the service decision visible. The case did not require choosing between dismissing customer convenience and celebrating revenue without its operating consequence. It required a defined account of the hours, contribution, added expense and uncertainty. A leader can decline an automatic expansion while continuing to examine a more selective offer, provided the next stage is described honestly and not presented as an achieved solution before it has been tested.