Every conversational-form vendor quotes a “3× completion rate” number. Nobody shows their working. We’re going to — specifically for Arabic forms, where the data behaves differently enough that the global benchmarks don’t transfer.
The numbers below come from 12,847 submissions across 241 published forms on Sahl/form between October 2025 and March 2026, with customer opt-in and all PII stripped before aggregation. Raw chart data is downloadable at the end of the post.
Methodology
We compared two modes of the same form, A/B split at the URL level for half of the forms in the dataset:
- Classic — page-based pagination, 3–6 questions per page.
- Conversational — one question per screen, auto-advance after answer, keyboard shortcuts.
A submission counts as “completed” when the respondent reaches and submits the final page. Everything else is a drop-off, attributed to the last question the respondent attempted.
Headline
Across the full dataset, conversational Arabic forms completed at 62.4% versus 23.1% for classic — a 2.70× ratio, not quite 3× but within the range we’ve been claiming. What’s more interesting is that the gap widens with form length:
- Forms with 1–4 questions: conversational 71% vs classic 58% (1.22×)
- Forms with 5–9 questions: conversational 64% vs classic 28% (2.29×)
- Forms with 10–19 questions: conversational 58% vs classic 14% (4.14×)
- Forms with 20+ questions: conversational 49% vs classic 6% (8.17×)
The “3×” claim is fair on average; the real story is that conversational mode does the bulk of its work on longer forms, which is where classic mode collapses.
Three conditions where conversational wins most
- Mobile respondents. 78% of Arabic-form traffic is mobile. One-question-at- a-time is built for that screen; classic pagination is a compromise between desktop and mobile that favours neither.
- Single-choice / rating / NPS question types. Auto-advance turns these into a single tap — the median time per question in conversational mode is 2.1 seconds for these types versus 6.8 seconds for classic. Shorter time per question, lower fatigue, higher completion.
- Forms with welcome screens. Adding a welcome screen lifts completion by a further 11 percentage points on average, regardless of mode — people commit more once they’ve clicked “Start.”
Where classic still beats conversational
Honesty point. Classic mode outperforms conversational when:
- The form is a quick-reply (1–3 questions). The conversational chrome is overhead.
- The respondent is on a desktop with a known identity (authenticated submitters, like internal staff forms). Completion ceiling is already high; conversational adds friction.
- The form contains a large file upload as the primary action. Full-screen focus on a file field feels slower than a classic layout where the field sits alongside other context.
What this means if you’re building a form
A rule that holds in our data: if the form has more than four questions or is aimed at mobile respondents or an anonymous audience, conversational mode is the default. If it doesn’t, classic is fine and sometimes preferable. The mode switch on Sahl/form is one click, so there’s no reason to over-commit either way.