Building software around messy real-world data teaches you humility quickly.
The moment you think:
"There's no way anyone would format a product like THAT."
You find three vendors doing exactly that.
One thing we're deliberately NOT trying to make PeptidePriceMatch:
A leaderboard declaring one vendor the winner.
I'd rather expose the underlying information and let people decide what matters.
Median price might be one of the most underrated market statistics.
The cheapest listing tells you the floor.
The most expensive tells you the ceiling.
The median tells you where the market actually sits.
Here's a rule I wish every price comparison followed:
Never substitute another size just because the exact size isn't available.
Missing data is better than wrong data.
A $100 product with 20% off isn't a $100 offer anymore.
It's $80.
Seems trivial.
Apply that across thousands of listings and suddenly discount logic becomes a core part of price comparison.
There's an important difference between:
CHEAPEST PRODUCT
and
CHEAPEST COMPARABLE OFFER.
A price comparison platform needs to understand the second one.
Something I'm excited about as PPM accumulates data:
Eventually the question isn't just:
"What does this cost?"
It's:
"What DID this cost?"
Price history makes the market much more interesting.
One product.
10mg.
20mg.
30mg.
40mg.
Those aren't four buttons attached to one price.
They're four separate offers that need to be tracked independently.
More listings β better data.
More CORRECTLY MATCHED listings = better data.
That distinction has become increasingly important as PeptidePriceMatch has grown.
Imagine:
Vendor A: 10mg β $40
Vendor B: 20mg β $45
Sorting those together and declaring Vendor A "cheaper" tells you almost nothing.
Comparable first.
Price second.
Building PPM has made me appreciate boring data fields way more than I ever expected.
Strength.
Unit.
Formulation.
Quantity.
Availability.
Get the boring stuff wrong and everything downstream is wrong.