A saved lot should not sound like reserved stock

The saved-lot action was easy to use. A buyer could keep an interesting listing, return to it later and discuss it with a project colleague. We had designed the control as a convenience in discovery. In a sixty-person task study, thirty-nine participants believed it would also hold the stock while they considered the purchase. Twenty-one understood it as a shortlist. The problem was not that users could not find the feature. It was that completing it created two different beliefs about what the marketplace had committed on their behalf.

For a buyer planning around a particular surplus lot, that difference could matter. They might arrange a site discussion, seek a specification decision or coordinate collection while assuming the material was waiting for them. The seller might meanwhile be considering another enquiry through a different route. The platform had recorded interest without obtaining a stock commitment. A note buried elsewhere in the process would not necessarily undo the ordinary meaning a customer took from the action. We needed to examine that meaning as part of the product, rather than treat it as a misunderstanding outside our responsibility.

Test the belief formed after the action

Usability research can become too narrow when it stops at successful completion of a task. A participant who finds the save control and uses it quickly may appear to have succeeded. The more consequential question is what they believe will happen next. Will the seller know, will the stock remain available, has a request been sent and is any further confirmation required? Those expectations connect the interface to the service. If the product measures only speed or clicks, it can miss a decision that has become less reliable precisely because the action feels so easy and final.

The commercial team preferred a short flow with a qualification explaining that availability could change. That approach preserved the convenience of the existing design and avoided introducing more seller coordination. Danish argued that the distinction should appear at the point where the buyer formed the expectation. I agreed. We spent ₹31 lakh on a redesign that separated a personal shortlist, a request for a hold and a seller-confirmed hold. The intent was to make the stages understandable in ordinary terms, without asking the buyer to infer the service state from an internal status code or an ambiguous reassuring message.

The study did not establish that thirty-nine of every sixty real buyers were misled. It was a task exercise that revealed a consequential interpretation worth addressing. We also did not assume that the twenty-one participants who understood the shortlist had the same needs as those who wanted a hold. A buyer can understand a feature perfectly and still need a different service. The design question therefore included whether a confirmed hold could be supported, under what conditions and with what responsibilities, rather than simply changing the label and declaring the underlying customer requirement solved.

The seller must be able to make the commitment

A platform cannot create a credible reservation by describing one more confidently. The seller needs to confirm the relevant quantity and the arrangement's stated period and conditions. Seller operations and legal colleagues retained that process and the actual terms. Product's responsibility was to represent the commitment accurately and help both parties understand what remained to do. We did not treat an expression of interest as authority to remove stock from another discussion, nor assume that every listing could support the same hold arrangement. The service had to earn the state the interface wanted to display.

This introduced friction. Some buyers would have to wait for a seller response, and some sellers would decline. The earlier flow made those possibilities less visible by leaving the meaning unresolved. The redesign did not create the uncertainty; it gave the existing uncertainty a place in the experience. That distinction mattered when colleagues judged the new route against the speed of the old one. A fast action that leaves two parties with incompatible expectations is not necessarily a simpler transaction. It may merely postpone the explanation until changing course is more expensive for the buyer's project.

A confirmed hold still needed proportionate language. The product should identify what had been confirmed without implying that every practical or contractual risk had disappeared. Quantity, timing and conditions belonged in the approved arrangement. Questions about rights, remedies or a disputed commitment remained for the responsible owners and the actual facts. I did not want product copy to become an improvised legal interpretation. Its job was to make the service state intelligible and direct the user toward details needed for the next decision, with no broader promise than the business could support.

Pending is a real stage of the service

The later review recorded fifty hold requests. Thirty-four received seller confirmation, ten were declined and six remained pending at the cutoff. These counts were useful because they showed the stages customers had actually reached. They did not describe fifty sales, a conversion improvement over the old design or the final outcome of every request. A confirmed hold could still be followed by further project decisions. A pending request was not automatically abandoned, and a declined request did not by itself explain whether the buyer found another suitable lot elsewhere in the marketplace.

We kept those distinctions in the management account. Otherwise, the product team could have reported every request as a successful reservation while operations still waited for six responses and knew that ten sellers had declined. The same clarity required for the user was required for leadership. A funnel becomes less useful when its stages borrow the meaning of later commitments. The result of our programme was a clearer service path and a defined set of observed request outcomes. It was not a demonstrated increase in revenue or a claim that all availability disputes had ended.

I would now ask the team to write down the belief each consequential action should create before selecting its label. That belief should be tested with buyers and checked against what sellers and operating colleagues can actually do. The work can reveal where a new service is needed, where a simpler explanation is enough or where the company should decline to imply a commitment at all. It also gives engineering a more precise product requirement: the state is not just a value to store, but an account of something that has happened between people with different responsibilities.

For directors, the question is how a marketplace's interface changes the commitments customers believe they hold. A board can examine whether product measures reflect those commitments and whether management understands the unresolved stages of the service. It need not choose button labels or review individual hold requests. It can require the enterprise to connect the screen, the operating process and the reported result. Our saved-lot action became clearer when we stopped asking one convenient feature to mean both remembrance and reservation. The extra stage was justified because the customer's project needed to know the difference.