Match Accuracy
Juice worth the squeeze
Winning in wholesale requires finding product opportunities that other sellers can’t find. Not only do you get better pricing with less competition, but deeper search allows you to make bigger buy orders, strengthen supplier relationships, get discounts, and drive more to the bottom line.
It is critical to get an edge against the competition, and to do so efficiently. Thus, you need recommendations that can be reviewed efficiently, and that is why our matching accuracy is critical to your success.
What accuracy should you expect from the data?
Our data leverages the industry-leading Keepa API, which is trusted by sellers as the most reliable source of Amazon data. We also are constantly keeping our calculations up to date for the most reliable profitability data. However, the profitability is only trustworthy if the conversion factor and match are correct. Both of these inputs are AI-generated and must be checked.
What accuracy should you expect from the AI?
When your supplier sheet has a normal amount of profitable opportunities, you should find:
- 90%+ of code-match recommendations are matches
- 25%+ of text match recommendations are matches
- 70%+ of conversion factors are accurate.
However, a sheet with zero profitable opportunities, would only show you the mismatches and the conversion factor mistakes, so you might see an accuracy score of 0%!
We break down this logic in the next section.
Post-Filter Accuracy
If you’re familiar with our scanning format, the opportunities tab shows products that have been filtered based on profitability. Unfortunately, whenever there’s a mismatch or conversion factor issue, these “false positive” products tend to make it through the profitability filter. For example, an iPhone 15 case that incorrectly matches with an iPhone 15 listing would be highly profitable.
In the image below, we depict how our 93% match accuracy score can become ~84% or ~50% after the profitability filter is applied.
Among all the matches, when we have clean data, we have found in through testing that around 26 of 28 products will be accurate (depicted by the shaded circles).
However, in scenario A above, if 20% of products are profitable, then you’re now left with 6 true opportunities. Meanwhile, roughly half of mismatches are profitable, so you have 1 incorrect opportunity, leaving 6/7 correct.
In scenario B, which is more realistic, if only 2% of products are profitable, then you end up with 1/2 correct in the opportunities tab.
To add to the challenge, your experience will show a lower accuracy if the data is dirty or if the supplier sheet does not have good pricing. For this reason, we have adopted the 90-25 standard to set expectations.
Conclusion
Every tool requires that sellers verify recommendations. We save you the time of vetting bad opportunities with our volume calculations and our conversion factor estimates.
Ultimately, you’ll be increasing your potential for gain and hopefully decreasing the time and cost it takes to get there.
If for whatever reason, you find a lower scanning accuracy than expected, please send us your feedback and we’ll refund your scan.
