There is a sequencing error that appears in most commercial turnaround programmes we are asked to review, and it is almost always made for defensible reasons. Demand generation is visible, measurable and produces movement within weeks. Conversion engineering is invisible, produces no immediate topline, and requires capital. So demand comes first.
The consequence is predictable and expensive. Acquisition spend delivers qualified visitors to a checkout that cannot serve them; those visitors leave, complete their booking through a marketplace that can, and the property pays twice — once for the click and once for the commission.
A property that could not take a corporate booking
A 95-key all-suite property on the Bengaluru Outer Ring Road corridor sat inside its own addressable market. The corporate demand it was designed to serve worked in the buildings around it. It was capturing almost none of that demand on contract, and 82% of what it did capture arrived through marketplaces.
The booking engine, unchanged since 2017, abandoned 88% of sessions that reached the payment step. Session recordings across 4,200 sessions identified the causes without ambiguity.
| Barrier | Sessions affected | Abandonment rate at barrier |
|---|---|---|
| Mandatory account creation before checkout | 3,712 | 54.1% |
| Card-only payment, no UPI or net banking | 2,940 | 38.7% |
| No GST detail capture at booking | 1,106 | 71.2% |
| 11 form fields before payment step | 4,200 | 22.4% |
| Session timeout under 8 minutes | 860 | 61.5% |
Payment-step abandonment causes, 4,200 recorded sessions, 95-key corporate property
Each of these is individually a solved problem in Indian e-commerce and has been for several years. Collectively they made the direct channel unusable for exactly the guest the property most wanted: a corporate traveller booking on a phone, needing a GST invoice, paying by UPI, and unwilling to create an account for a two-night stay.
What happens if you buy demand first
It is worth working through the counterfactual, because it is the path most programmes take. Suppose the property had begun with corporate contracting and paid acquisition, and left the engine for a later phase.
The contracting workstream would have succeeded on its own terms — corporate rates loaded, accounts signed, travel management systems updated. Those corporate travellers would then have attempted to book. Eighty-eight percent of them would have failed at the payment step, and would have completed the booking through the marketplace or the travel management company's own channel instead.
The property would have paid to create demand, then paid commission to serve it, and the reported outcome would have been rising occupancy alongside deteriorating margin — a pattern almost always misdiagnosed as a rate problem.
The economics of fixing the engine first
The remediation was not technically ambitious. Checkout reduced from five steps to three. Guest checkout enabled, account creation made optional. UPI, net banking, corporate card and pay-at-hotel added. GST details captured at booking with automated invoice issue. Form fields reduced from eleven to five. Session timeout extended.
Payment-step abandonment fell from 88% to 41% within six weeks of deployment. Against the property's existing direct traffic — before any incremental acquisition — the recovered volume alone repaid the development cost in seven months.
| Metric | Before | After | Movement |
|---|---|---|---|
| Payment-step abandonment | 88.0% | 41.0% | -47.0 pts |
| Session-to-booking conversion | 0.8% | 2.9% | +263% |
| Mobile conversion | 0.4% | 2.6% | +550% |
| Bookings with GST capture at booking | 0% | 64% | +64 pts |
| Average booking value | ₹8,940 | ₹10,120 | +13.2% |
Direct channel performance, pre and post engine remediation, constant traffic volume
Average booking value rose 13.2% as well, which surprised the client. The explanation is that the prior checkout selected for the most determined guests — typically short, simple, single-room bookings. A functioning checkout admits the longer, more complex, higher-value bookings that had previously been abandoning.
A rule for sequencing
We now apply a simple gate in commercial programmes: no incremental acquisition spend is committed until payment-step abandonment is below 60% and mobile conversion is within one third of desktop. These are not aspirational targets. They are the thresholds below which acquisition spend reliably subsidises intermediaries.
Hard qualifiers for the Indian market
- UPI at checkout. On a mobile-majority booking profile, card-only payment is a structural exclusion, not an inconvenience.
- GST detail capture at the point of booking, with automated invoice issue. In reimbursable corporate travel this is a qualification criterion, not a finance detail.
- Guest checkout without mandatory account creation. Account walls are the single largest abandonment cause we measure.
- Mobile LCP under 2 seconds. Every additional second of load time on a mobile-majority profile is measurable lost revenue.
This analysis draws on engagements led by the author. Findings are anonymised at client request; underlying figures are taken from client property management, channel manager and advertising platform records.
