Variants and the user's own data
Once a study has run, change one input, such as a price, a rate or a volume, and Scnro runs a variant with everything else fixed. Then compare up to four reports side by side.
Your advisor will ask you for
- A study with a finished run.
- The input to change, and its new value.
- For a dataset figure: dataset, period and unit.
What you get
- A named variant that states what changed.
- A side-by-side comparison of up to four versions.
- A data quality report before a dataset is relied on.
Jobs covered
| Run what-if variants and compare them | |
|---|---|
| Build a variant from dataset figures | |
| Explore data and check its quality | |
| Make a synthetic dataset |
Details of each job
- Run what-if variants and compare them
- Change an input, rerun, and compare up to four versions side by side.
- Build a variant from dataset figures
- Bring a figure from a synthetic dataset into a variant, labelled as an assumption.
- Explore data and check its quality
- Look over a dataset and check its quality before relying on it.
- Make a synthetic dataset
- Design and generate a sample dataset for your problem, with every number tied to a source and its limits stated.
How it works
Who does what. Prices and consent: Using Scnro.
- 1 · YouStudyA variant needs a study that has run once.
- 2 · YouChangeOne price, rate or volume.
- 3 · ScnroRerunEverything else stays fixed.
- 4 · ScnroCompareUp to four versions on the same basis.
- 5 · YouDecideSee which assumption moves the answer most.
Example prompts
Copy one and change the details.
Use Scnro to rerun my second-shop study with rent 15% higher and show what changes.
Use Scnro to compare the original study with the higher-rent and fewer-cups variants.
More example prompts
Use Scnro to take the average daily transaction value from the payment dataset into a variant.
Use Scnro to check the quality of the synthetic dataset my team generated.
Use Scnro to run three variants of my study (rent +15%, cups −10%, both) and compare all four.
Use Scnro to explore the insurance dataset and tell me which figures I could use.
Tip
Change one thing per variant, name it (“rent +15%”) and compare at most four.
Troubleshooting
A variant cannot be created.
It needs a study with a finished run.
The comparison leaves a figure out.
The versions measure it differently; Scnro keeps them apart.
The variant changed more than asked.
Ask Scnro to list what changed and revise it.