> For the complete documentation index, see [llms.txt](https://docs.enlyze.com/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.enlyze.com/en/production-optimization/analysis/product-analysis.md).
# Product analysis
One product, all its production runs: consistency and specification adherence across recent batches.
Product analysis inverts the question [Production run analysis](/en/production-optimization/analysis/production-run-analysis.md) asks. Instead of picking individual runs, you pick a product and see how evenly it has run across its recent batches.
Useful for reading off the effect of a process change: did the runs after it track the target more tightly than the ones before. Equally, when a customer complains and you want to know whether the batch involved stands out.
Pick a product, set the scope, confirm the variables
## Prerequisites
* A machine with booked production runs
* At least two runs of the same product, otherwise there is nothing to compare
* Optionally a [setup sheet](/en/production-optimization/setup-sheet.md) for the product, see **Data source** below
## How it works
{% stepper %}
{% step %}
## Pick a product
The **Product** table lists the products made on this machine with **Product Description**, **SOP**, number of **Orders** and **Last order**. The **SOP** column shows whether a setup sheet exists.
{% endstep %}
{% step %}
## Set the scope
Under **Order selection**, choose **Last 5** or **Last 10**. Runs from the past two years are considered, capped at ten to keep the data volume manageable.
{% endstep %}
{% step %}
## Check the variables
Under **Variables to analyze**, the evaluated parameters appear as chips. The preselection comes from the machine's setup sheet template. You can adjust it, or load a [variable group](/en/production-optimization/analysis/variable-groups.md) with **Load group**.
{% endstep %}
{% step %}
## Load the data
Select **Load data** at the top right.
{% endstep %}
{% endstepper %}
## Data source: statistics or setup sheet
The analysis needs a tolerance band to check against. Where that band comes from decides what the percentages mean.
| Choice | Band | What it tells you |
| -------------------------- | --------------------------------------------------------------------------------- | ---------------------------------------------------------- |
| **Variables** | Derived from the runs themselves, mean ± 2σ. The view labels this **Statistics**. | **Consistency**: did the runs behave alike |
| **Setup data sheet (SOP)** | The limits defined in the setup sheet | **Specification adherence**: did the runs stay within spec |
The difference matters. With **Variables** as the basis, every run can score 100 % and still sit outside specification, because the band came from those same runs. Only a setup sheet measures against a real target.
## Metrics
Above the matrix, five tiles cover the selected product: **Production runs**, **Total quantity**, **ø Throughput**, **ø Scrap** and **ø OEE**, each with a trend line and the change against the previous period. Top right shows the average consistency across all parameters and the run marked as the **Golden Run**.
## Conformance matrix
Parameters as rows, runs as columns. Colour shows how much time stayed in band.
Rows are parameters, columns are individual runs. Each cell shows the share of production time the parameter stayed within band.
| Colour | Meaning |
| ------ | -------------------- |
| Green | 97 % or more in band |
| Yellow | 85 to 97 % |
| Red | below 85 % |
Above the matrix, the **Before** and **After** rows show the product made before and after, and what the transition looked like. **Direct** means the run followed without an interruption. Otherwise the cell gives the number of downtimes in between and their total duration, such as `4 downtimes · Σ 2.4 h`. That surfaces start-up problems inherited from the transition.
To the left of the run columns, each parameter shows its average across all runs with unit and target value. The final row summarises consistency per run.
## Parameter detail
Below the matrix you select one parameter and see its batch-to-batch progression as a boxplot per run, laid over the band. The table beside it gives mean, σ and consistency percentage per run.
A single red boxplot among otherwise green values points at an event in that specific batch. Wide spread across every run points at a process problem.
Next to **ø Consistency** sits the parameter's **Spread**. The legend under the plot names the band in use, the mean, and what box and whiskers represent (box = q1 to q3, whiskers = 95 %).
## Timeseries comparison
Below the parameter detail you can compare two individual runs directly: pick **Production order 1** and **Production order 2**, and the selected parameter's full trace is overlaid for both, aligned at run start. Detected downtimes appear as shaded areas.
Two runs overlaid, aligned at run start
The legend under the plot gives mean and σ per run. **Open in production run analysis** at the bottom right carries the same selection into [Production run analysis](/en/production-optimization/analysis/production-run-analysis.md).
{% hint style="info" %}
All statistics cover actual production time. Samples inside detected downtimes are excluded, exactly as in [Production run analysis](/en/production-optimization/analysis/production-run-analysis.md).
{% endhint %}
## Common questions
Why is there a dash in the SOP column?
No setup sheet exists for that product. The analysis still works, but only with **Statistics** as the basis. For specification adherence, create a [setup sheet](/en/production-optimization/setup-sheet.md) first.
Why do I see fewer runs than expected?
Three limits combine: runs from the past two years only, ten at most, and only runs on the selected machine. The same product on another machine appears there.
What does direct mean for a transition?
There was no detected interruption between the previous run and this one. A transition that is not direct had downtime in between, which can affect the run's start-up values.
## Related topics
* [Batch record](/en/production-optimization/analysis/batch-record.md) to document a single run.
* [Golden Run](/en/production-optimization/golden-run.md) explains how the most productive run is determined.
