Stratum vs Minitab

Minitab has been the default statistics package in Six Sigma training rooms for decades. Its control charts, capability studies and Assistant are good enough that many engineers have never used anything else. Before you renew, though, three facts are worth stating plainly:

• The desktop app is Windows-only. Minitab's system requirements list the current desktop release for Windows 11 on 64-bit x86 processors, with ARM not supported. On a Mac, your options are Minitab's web app in a browser or Windows in a virtual machine.
• It is subscription-only. Minitab sells annual, named-user plans. The bundles on its store that include Minitab Statistical Software start at $2,394 a year, and no perpetual license is offered.
• Most SPC work is standard SPC: I-MR and X̄-R charts, attribute charts, run rules, and a Cpk you can defend. You don't need a quality-engineering platform for that.

Stratum is a native app for Mac and for Windows, sold once. Control charts and process capability are in Professional ($249), and Standard ($119) covers the everyday statistics toolkit.

What Minitab is genuinely good at

Minitab's real advantage is quality-engineering breadth: design of experiments, measurement systems analysis (Gage R&R), reliability and survival analysis, acceptance sampling, transformation-based capability for non-normal data, and specialist charts such as Laney P′/U′ and the rare-event G and T charts. The Assistant walks a green belt through each method, and there is a large ecosystem of training material built around it. Stratum is a general statistics and charting app that does standard SPC properly; DOE and MSA sit outside what it sets out to do.

Control charts, head to head

Stratum ships six control-chart families under Charts ▸ Control Charts: I-MR Chart, X̄-R Chart (with an R or S panel), P Chart (p or np), C Chart (c or u), CUSUM Chart and EWMA Chart. These are the controls you'd look for coming from Minitab:

• Phases. Set a Phase column and the limits are estimated from a Baseline phase, then frozen and extended across the rest (the Phase I / Phase II convention). Alternatively, each phase can get its own limits. Run rules restart at every phase boundary.
• Exclude Point from Limits. Right-click a known special cause to keep it out of the estimate. It stays on the chart, drawn hollow, and is still judged against the limits and the rules.
• K Sigma. Draw limits at 3σ, 2.5σ, 2σ or any width you type. P, NP, C and U charts also accept a known p, c̄ or ū as Custom limits.
• Rules. All eight Nelson rules, with a Western Electric Only shortcut (rules 1, 2, 5 and 6) and Show Zones. Rule 1 alone is the default, as in Minitab and JMP. Zone rules apply to the location panel only, never to a range or moving-range panel, where they would fire spuriously.
• Points you can name and trace. The Labels control names points from your own columns, for example only the out-of-control subgroups. Reveal in Datasheet on any point takes you to its row.

Capability with an honest error bar

Analyze ▸ Process Capability reports Cp, Cpk, Pp, Ppk, CPL/CPU, Cpm and Z.bench, plus observed and expected PPM, under a capability histogram with your spec lines. In Stratum 1.5, every index also carries confidence intervals at 90%, 95% or 99%. Cp and Pp use the chi-square interval. The one-sided indices use the exact noncentral t, and Cpk and Ppk take the interval of whichever limit is nearer. That is the answer to whether a Cpk of 1.4 from thirty parts really clears 1.33. Often it doesn't. Process capability, explained covers the indices themselves.

The rest of the report covers what a Minitab user would check:

• The report flags data that fails an Anderson-Darling normality check.
• An optional 1.5σ Shift adds the Six Sigma convention's long-term Z and DPMO.
• Distribution offers Lognormal, Weibull, Exponential and Best Fit, computed by the ISO 22514-2 percentile method.
• The conventions follow the ones Minitab documents: for subgrouped data the within sigma defaults to the pooled standard deviation with c₄ unbiasing, and the overall sigma is the plain sample standard deviation.

Where Stratum chooses differently, the validation record says so. The interval bounds are checked against R's qcc and SciPy's noncentral t.

Side by side

Try it on a real process

Download the piston-ring dataset (Montgomery's 40 samples of 5, spec 74 ± 0.05) and build an X̄-R chart with Phase set and Baseline on Phase I. Then run Process Capability on the same column. In Control and Capable walks through the whole study, and What Is a Control Chart? covers the basics. Stratum opens .csv, .xlsx, SPSS .sav and Stata .dta files directly. To bring a Minitab worksheet across, save it as Excel or CSV first.

Download Stratum Stratum vs JMP →

Frequently asked questions

Does Minitab run on a Mac?

Not as a desktop app. Minitab's own system requirements list the current desktop release for Windows 11 on 64-bit x86 processors only. On a Mac you use Minitab's web app in a browser, or run Windows in a virtual machine. Stratum is a native app for Mac and for Windows.

Is there a Minitab alternative with a one-time price?

Yes. Minitab is sold as annual named-user subscriptions. Stratum is a one-time purchase: Standard is $119 and Professional, which includes control charts and process capability, is $249.

Which control charts does Stratum have?

Six families: I-MR, X̄ with a Range or StdDev panel, P and NP, C and U, CUSUM, and EWMA. They support phased limits, excluding a point from the limit estimate, a custom K Sigma width, known (Custom) limits, and all eight Nelson rules with a Western Electric shortcut.

Does Stratum give confidence intervals for Cpk?

Yes. Cp, Cpk, Pp and Ppk carry lower and upper bounds at 90%, 95% or 99%: chi-square for Cp and Pp, and the exact noncentral t for the one-sided indices, with Cpk and Ppk taking the interval of the nearer limit.

What does Minitab do that Stratum does not?

Design of experiments, measurement systems analysis such as Gage R&R, reliability and survival analysis, acceptance sampling, the Assistant, and specialist charts such as Laney and rare-event charts. Stratum is a general statistics and charting app with standard SPC, not a quality-engineering suite.

Will Stratum's capability numbers match Minitab's?

Stratum follows the conventions Minitab documents: the within sigma for subgrouped data defaults to the pooled standard deviation with c₄ unbiasing, and the overall sigma is the plain sample standard deviation. Every convention, and every place Stratum deliberately differs, is published in its methods guide and validation record.