Which is the best statistical software for Lean Six Sigma or statistical problem solving?
This is a question that quality professionals often ask, when they are starting to foray in to any kind of structured improvement program. Irrespective of the forum, most often, the question is about Minitab vs xxx. This indicates the popularity of Minitab. So I decided to do a series of posts comparing different statistical software available in the market, one at a time. The first one in this series is going to be Minitab vs SigmaXL
Minitab vs SigmaXL
While there are many good statistical software - both paid and free or opensource, Minitab and SigmaXL are one of the most popular in the Lean Six Sigma professionals community. Here is a summary of my evaluation of the latest versions of the software - Minitab 18 and SigmaXL 8.0.
| SL | Features | Minitab 18 | SigmaXL 8.0 |
|---|---|---|---|
| 1 | Software Type | Stand alone | Microsoft Excel Add-in |
| 2 | Individual User Licence | Yes | Yes |
| 3 | Concurrent user licence | Yes | No |
| 4 | Primary User Base | Business Users + Statistical and Six Sigma experts | Statistical & Six Sigma Experts |
| 5 | Popularity | 90% of Fortune 100 companies use Minitab Statistical Software | Mostly used by smaller companies, and expert consultants. |
| 6 | Famialirity | Needs initial familiarity time | Excel based, so excel users feel familiar |
| 7 | Operation System | Windows Only | Windows, Mac |
| 8 | Learning Curve | Smoother - users with relatively less experience in statistics can start using in-built Minitab Assistant feature to decide which tool to use | Steeper - requires thorough knowledge of statistics for use, and guidance is limited to self-research. |
| 9 | Help Documentation | Exhaustive documentation, Online as well as inbuilt/offline point of use help, with brief descriptions, explanation of statistical concepts, how-to instructions, sample data, explanations and tutorials | Limited to how-to documentation in a downloadable PDF workbook |
| 10 | Pricing | Costlier for individual users, but can be cost effective considering concurrent (can be installed in unlimited computers) user licensing option | Cheaper for individual user licence, but concurrent user licencing option doesn’t exist |
| 11 | Graphical Tools | Rich Graphical reports that can be saved as projects, and exported /copied to Microsoft Word and PPT | Graphical reporting limited to features available in Excel |
| 12 | Auto updation of Graph | Many graphs, including control charts can be updated automatically (or manually) by adding new data | Analysis needs to be re-executed |
| 13 | Pre-analysis of data | Uses many pre-analyses of data to check adequacy of sample size, dispersion of data, stability of data, and auto-recommends statistical methods (for example recommends p-chart or laney p-chart automatically) | No pre-analysis or recommendation feature. Give results as per statistical formulas, based on options selected by the user |
| 14 | Statistical Tools | Comprehensive - basic to most advanced tools are available | Most basic tools are available. Limited number of advanced tools are available. |
| 14.1 | Hypothesis tests | Yes | Yes |
| 14.2 | ANOVA | More than 2 way/factors possible, including ANOVA general linear model | Only up to 2 way is possible. ANOVA general linear model, and multi-way ANOVA is not available |
| 14.3 | Corelation and Linear regression | Yes | Yes |
| 14.4 | Power and sample Size | Yes | Yes |
| 14.5 | Control Charts | All types of control charts avaailable | Limited types. Cusum, EWMA not available |
| 14.6 | Design of Experiment | Yes, including general full factorial - capable of designing and analyzing simple to complex experiments | DOE available for up-to two level and fractional factorials. Complex design of experiments not supported |
| 14.7 | Response surface method | Yes | No |
| 14.8 | Reliability Analysis | Comprehensive reliability analysis tools available | Limited to Weibull analysis |
| 14.9 | Non-normal data analysis | Comprehensive options of non-normal distribution | Limited to converting non-normal data to normal using Box-Cox transformation |
Verdict:
As long as advanced statistical analysis is not required, and you and your team are confident about the tools and options to use, SigmaXL will suffice your needs, and save some money.
However, if you or your team members are primarily business users, who want to use statistics as a tool to solve business problems or get deep insight into your data, Minitab should be preferred over SigmaXL.
