Research · GIA Operator Outlook · v1.3 · Pilot
Quarterly SME confidence across six African markets
The Outlook is a quarterly, forward-looking reading of the operating confidence of small and medium businesses (5–500 people), measured country by country, independently. It borrows the diffusion method of the PMI, but it is not a PMI: the headline is built from what operators expect and intend for the quarter ahead. Every number is published with its confidence tier and sample. Where a country's sample isn't strong enough, we say so — never a fabricated figure.
Six markets, each on its own confidence tier. Nigeria fields from the six-year Naijabusinessguy operator panel; the other five field through country desks and panel partners currently being recruited. A country never borrows another's credibility. Fielding window: final four weeks of Q4 2026.
n too low to publish · Naijabusinessguy operator panel — fielding open.
n too low to publish · Field & panel partner — recruiting.
n too low to publish · Field & panel partner — recruiting.
n too low to publish · Field & panel partner — recruiting.
n too low to publish · Field & panel partner — recruiting.
n too low to publish · Field & panel partner — recruiting.
Every score runs 0–100. We do not attach labels like "strong" or "weak" to any range — those would need years of data we do not yet have.
As many operators answered 'up' as 'down'. No net change.
More operators answered 'up' than 'down'.
More operators answered 'down' than 'up'.
Each country publishes at the tier its actual valid sample supports that quarter. These are publication thresholds — the minimum number of valid responses before we publish — not a claim of national representativeness.
Headline, all readings and the precision benchmark are published.
A real but provisional reading, clearly flagged, with sample size stated.
No number. Reported honestly as 'coverage building, n too low to publish'.
Every respondent, in every country, answers the same six forward-looking questions on one 3-point scale. Four build the headline. Two are published as separate gauges — credit conditions and cost pressure are conditions of the environment, not the operator's own plan, so they sit outside the headline.
The Outlook uses the diffusion method — the same family as the PMI. For each item, in each country:
The Cost Pressure Gauge (D5) is not inverted: a higher score means more operators expect costs to rise. The Credit Conditions Gauge (D4) is scored only among businesses seeking credit, with the share not seeking stated; a higher score means credit is expected to get easier. Equal weighting is used because anyone can check it; whether other weights give a better early signal is tested before a country's eighth quarter, and any change is a disclosed version.
A margin of sampling error only has a statistical basis when every business had a known chance of random selection. Our panels are recruited from our own audiences and partners, so that condition is not met. Printing "±5%" would suggest a certainty we do not have.
Instead, every published figure carries a Random-Sample Precision Benchmark — how far a single item score could move by chance if the sample had been random. It is stated in points, never as a percentage, and it is a benchmark for comparison, not a confidence interval for our sample.
The Outlook does not estimate the share of all SMEs in a country, or in Africa, that are confident. It measures the balance of responses among the valid operators who took part in that quarter's sample. Recruitment is not random, so a figure should not be read as a population estimate without looking at who was surveyed. It is not a margin-of-error poll, and it does not cover micro businesses of 1–4 people, whose answers are kept separate. Six markets are a deliberate cross-section, not a sample of Africa, which is why we never publish a single continental figure.
Integrity
A country at "Building" is always preferable to a fabricated figure. Until independent methodological review is complete, releases carry a pilot label. Figures and commentary are kept apart: the figure block is published exactly as produced, and editorial interpretation never changes it. We never quietly revise a past number — a correction, if ever needed, is published as a correction. The full versioned methodology is published so any figure can be traced to exactly how it was produced.