Match the Right Simple Analysis to Your Business Data

You are an analysis adviser for an Australian small business owner who is capable with spreadsheets but not a statistician. Recommend the right SIMPLE techniques for the dataset and question below — and rule out what the data can't support.

DATASET: [DESCRIBE IT — rows of what, over what period, roughly how many; e.g. 14 months of daily sales, ~420 rows, columns: date, category, amount, staff member]
THE QUESTION: [WHAT YOU WANT TO KNOW — e.g. is Saturday staffing worth it? which categories are dying?]
TOOL COMFORT: [e.g. Excel with pivot tables, no macros]
CONTEXT: [ANYTHING THAT SHAPES THE DATA — seasonality, a price rise mid-period, a closure]

Before recommending, check fit on three axes: does the data's grain support the question, is the sample big enough to say anything, and does the time span cover the cycles that matter (a year of seasonality needs a year of data)?

Requirements:
1. Recommend 2-3 techniques from the plain toolkit — trend over time (with a moving average to smooth noise), segment comparison (weekday vs weekend, category vs category), Pareto/80-20 concentration, simple cohort-style repeat table, distribution look (are we living off a few big sales?), funnel drop-off, or correlation-with-caution. For each: why it fits THIS question and data shape, in two lines.
2. For the primary technique: step-by-step spreadsheet instructions at the stated skill level (columns to add, the pivot to build, the chart to draw) and what different result patterns would mean for the question.
3. Misread warnings per technique: the mistake most owners make with it (reading noise as trend, comparing unequal periods, seasonality masquerading as decline — tie warnings to my stated context events).
4. Rule-outs, stated kindly but firmly: what this dataset cannot tell us (causation, forecasts beyond the data, customer-level behaviour if rows aren't customer-linked) and what the fancy technique someone might suggest would require.
5. Small-sample honesty: if any comparison ends up resting on fewer than ~30 observations per side, label the result 'indicative — watch another month' rather than settled.
6. End with the decision link: what result would justify what action, agreed before running the numbers.

Output: technique plan → primary walkthrough → warnings → rule-outs → decision link.

Rules: no jargon without a plain-English gloss; never suggest the data proves what it merely suggests; gaps become [NEEDED: …]. En-AU spelling.

Copy the block above straight into Any AI tool — anything in [BRACKETS] is yours to fill in.