A data analyst looking at a column of figures asks nearly the same questions as a person facing a number series: what is the step between neighbours, is the change constant or growing, and is there a second pattern hiding in the first. The numbers are messier, but the questions are the same.
Linear or exponential: the gap-and-ratio check
Take two products. Product A gets 100, 115, 130, 145 sign-ups in its first four weeks, adding 15 a week. Product B gets 100, 115, 132 and 152 (rounded), growing by about 15 per cent a week. The first step is identical, which is what makes it easy to confuse them.
The difference shows by week eight. A reaches 205, and B reaches about 266. Compute the gaps for A and they stay at 15; compute them for B and they keep growing. Compute the ratios for B and they stay near 1.15. That check is exactly the pair of questions from a number series test.
Differencing: finding what changed
The standard time-series move of taking the change from one period to the next is the gap row in a series question. It strips out a steady trend and leaves the part that needs explaining.
Do it twice and you find out whether the change itself is steady, the second-difference idea from the series tests.
Seasonality: two series woven together
Quarterly figures often run 40, 65, 44, 70, 48, 75. Read straight through they zigzag. Split them and the structure is clear: the odd positions (40, 44, 48) rise by 4, and the even positions (65, 70, 75) rise by 5.
The next figure is in the odd list, so it is 52, and the one after it is 80. In a real dataset the split would be by quarter or by weekday, and the rule would only hold roughly, but it is the same instinct.
Sanity-checking an output
The quiet use of series reasoning is checking results. A forecast that adds 15 each week and then jumps by 90 has broken its own rule. Spotting that, before anyone else does, is a large part of being trusted with numbers.
Where the analogy breaks
In a test, 100, 115, 130, 145 has a rule and the next term is 160. In a spreadsheet the same weeks might read 101, 114, 131, 144, and the question changes: which simple pattern is this roughly following, and how big is the departure from it?
The series instinct gets you to the candidate pattern fast. Deciding whether the departures are noise, an outlier or a real change in behaviour is statistics and domain knowledge, and no short test measures that.
What a series test does and does not show
The JobCannon Number Series test gives you a quick read on how comfortable you are with rule-hunting in clean numbers, split by steady steps, growing steps, two series and chained rules. It does not measure statistics, coding or judgement under uncertainty, and as an in-house tool it has no published norms, so use it as practice feedback only.
For reading charts and working with tables, the JobCannon Numerical Reasoning test is the closer fit, and the blog's numerical reasoning for data analysis piece goes further. Drill the basics on the number series practice page.