Is one app wrong when the scores differ?
Not necessarily.
Why Can the Same Food Get Two Completely Different Scores?
Published · Updated · Sláinte Editorial
You scan the same cereal in two apps.
One gives it a strong score. The other gives it an average one.
Which app is wrong?
Possibly neither.
Before treating the two numbers like competing measurements, it helps to understand where food scores come from.
First check that it really is the same food
This is the boring step, but it matters.
Check:
- barcode
- flavour
- pack size
- market
- ingredients
- nutrition
If one app has an older recipe, you are not really comparing two scoring systems anymore. You are comparing two different inputs.
The underlying data can differ
Imagine one database says a product contains 1g salt per 100g. Another has an old entry showing 0.8g.
Even if the scoring formulas were identical, the outputs could differ.
Missing fibre is another example. “Not declared” does not automatically mean zero.
Nutrition models can differ
There is no law saying every private food scanner must weight sugar, salt, saturated fat, fibre, protein and energy in the same way.
Different methodologies can therefore disagree while using the same nutrition table.
Ingredient rules can differ
One app may strongly penalise a particular additive. Another may focus on whether the additive is authorised and consider it in the wider product context.
Those are different methodological choices.
Processing can differ
One methodology might incorporate processing directly. Another might primarily use nutrient composition. A third might show NOVA separately rather than mixing it into its main score.
Again, those numbers are not automatically measuring the same concept.
Caps and category rules matter
Suppose an app decides that certain confectionery products can never receive its highest rating even if individual nutrient values are relatively favourable.
That cap changes the result.
Another app without that rule may score the same numbers differently.
A score is not a universal measurement
Weight is measurable in kilograms. Temperature is measurable in degrees.
A proprietary food score is different.
The score is produced by a method.
That does not make it meaningless. It means the methodology matters.
Same label factsMethod A / Method Bdifferent outputs
Can both apps be useful?
Yes.
One might be better at exposing raw data. Another might make the result easier to understand. Another may have stronger coverage of the products you buy.
The real mistake is treating a number as self-explanatory.
Why can the same app change a score?
There are three common reasons.
Product data changes
A recipe is updated or corrected.
Missing data is added
For example, fibre becomes available.
The methodology changes
A scoring system may be revised.
If an app changes its rules, transparent versioning helps explain why historical results changed.
What this means when you’re shopping
If two apps disagree, look underneath the numbers.
Check:
- Do they have the same product?
- Do the ingredients match?
- Do the nutrients match?
- Are they scoring the same concept?
- Is one using extra category rules?
That will usually tell you more than arguing over whether 71 or 78 is the “real” score.
How Sláinte helps
Sláinte’s methodology should remain public so users can understand what the result means.
Frequently asked questions
Can two apps use the same database and still disagree?
Yes. They may interpret the same data differently.
Does a higher score mean a food is safer?
Not automatically. Safety and nutritional scoring are different questions.
Why did my product score change?
The product record or the methodology may have changed.
Is there one official universal food score?
No private scanner score should be presented as a universal official health measurement.
The short version
- First make sure the apps have the same product.
- Product records can differ.
- Nutrition models can differ.
- Ingredient rules can differ.
- Processing rules can differ.
- Category caps can differ.
- A proprietary food score is an interpretation, not a physical measurement.
- Read the methodology and underlying label facts.