Gulf Market and Audience Research, Primary Not Panel

In the last four food and beverage studies I ran in Saudi Arabia and the UAE, between a quarter and two fifths of live consumer demand arrived in spellings the client's keyword tool reported as zero volume: Arabizi, Khaleeji dialect, and undiacritised Arabic that the tool folded into an unrelated intent. If your market sizing came out of Keyword Planner and a syndicated online panel, you have measured the English-speaking urban expatriate slice and labelled it the market. What the research answers: Three questions, and I will not pad the document with a fourth. Who is buying this category here, in segments defined by behaviour rather than by demographics you cannot act on. What stops the ones who nearly buy, named as a specific obstacle at a specific step, with the count of interviews it appeared in. And what would have to change for that obstacle to disappear, which is frequently not a marketing change at all, in which case the report says so on the first page rather than the last. Fieldwork: Twenty to twenty-eight interviews, conducted in the respondent's own language: Khaleeji, Egyptian or Levantine Arabic as the segment requires, English where English is genuinely the buying language rather than the researcher's convenience. Recruitment does not come from a panel. It comes from your churned customers, from your competitors' public reviewers, and from referral chains inside the community that actually consumes the category, because Gulf online panels skew hard toward English-speaking Dubai and reproduce the exact error the study exists to correct. Alongside the interviews: search demand rebuilt by hand across transliteration and diacritic variants; checkout testing on your funnel and two competitors' covering mada, Tabby, Tamara, Apple Pay and cash on delivery; and observation of buying that never touches a website, which in this region still means WhatsApp and Instagram direct messages carrying real volume in food, and Telegram in virtual assets. Each category gets its own pass. In virtual assets the audience question is usually which claims may be made to whom, so the interview set includes your compliance officer and the findings are read against the VARA marketing rulebook rather than treating regulation as somebody else's chapter. For nonprofits, giving here concentrates violently in Ramadan and inside the last ten nights, and the distinction between Zakat and Sadaqah changes both the ask and who may lawfully receive it. A donor study that treats Gulf giving as year-round philanthropy is wrong before the first interview. Where the sample stops: Twenty-eight interviews is not a representative sample and I will not present it as one. It is enough to find the obstacle and nowhere near enough to size a segment to a percentage point. Every number in the report is labelled by origin: fieldwork, platform data, or an assumption I made. The assumptions are collected on one page so you can attack them in a single sitting. No personas with invented names and stock photographs. No net promoter score. No focus groups, because in this region deference to the most senior person in the room produces agreement rather than information. I turn this down when: The decision is already taken and the research is meant to ratify it. Everyone can feel that by week two and the document is worthless afterwards. The real question is a pricing question. You want a pricing study, and I will name two people who do them properly. The timeline is under three weeks. Recruiting a sample that is not a panel takes ten to fourteen days on its own, and a fast version of this work is a slow way to be misled.