August 7, 2020

(Part #2) ML – Important decisions to be made

Misha Krunic @mkrunic

Let's continue introducing you to our new project - product matching assisted by Machine Learning (ML).

Before kicking the project off, we had to make some really important decisions regarding the project scope.

1.Language-specific or universal ML model?

Of course, one would like his solution to be as broadly applicable as possible.

  • Language-specific model would probably be more precise but would require training for each language individually.

  • And, preparing a training set, as you will see is a very difficult one
    As Price2Spy has clients from literally all over the world, we would need to cover at least 15 different languages, and some of them written in non-Latin scripts.

Decision: try to go for a universal solution, by all means

2. Industry-specific or universal ML model?

  • Price2Spy works with over 25 different industries. Preparing 25 training sets to build 25 different ML models seemed like a nightmare.

  • Again, the industry-specific model would probably be more precise but would require training for each language individually. And, preparing a training set which is representative enough, as you will see is a very difficult one

Decision: try to go for a universal solution, by all means

3. Matching accuracy

  • One thing that we have learned in 9 years in this business is that a wrong match is something that we cannot afford to have in Price2Spy. Wrong match => Wrong pricing decision. Our customers cannot have that => we cannot have that!

  • 99% matching accuracy is not sufficient. Even if it’s only 1% of wrong matches – how can the client know which 1% is wrong?

Decision: we’re striving for 100% matching accuracy

So, we have our 3 key ML matching decisions. On to the next task – preparing the training set!

Today's Top Milestones
  • 🚀 A New SEO toolkit is launched on ProductHunt!
    Today is Screpy day! Screpy launched on Product Hunt now. A website analysis tool we’ve been hard working on for months. Will it be interesting? How m
  • I was interviewed for a podcast for the first time
    Last week I was invited to be interviewed for Scottish Business Podcast - a relatively new podcast that covers Scotland-based businesses. The host was
  • $100 MRR!
    With a couple of new customers yesterday, ilo.so is now at $135 MRR. This is a big milestone for me. By selling a number of lifetime deals, ilo has al
  • A Proposals generator is now on ProductHunt!
    Satiurn Proposals launched on Product Hunt now. It's a fast and frictionless tool for create stunning proposals, we hope this could be bring to Satiur
  • Public API
    I'd like to announce that SaaSHub has a simple public API now. There are two endpoints as of now. In general, that should be sufficient for some basic
  • Reached 70 users in 3 weeks 🎉
    Hi all 👋 Today we reached the first 70 users which created a form with Tally and shared it to collect responses. We launched the beta of Tally 3 week
  • Submitted to ProductHunt
    I'm releasing the product on ProductHunt, after the initial beta release and sharing the product with some fellow indie hackers. Although there are ma
  • 3 weeks until a wide release beta!
    You may wonder, if you read the previous update, on why a nearly done MVP, is still awaiting to a Beta release. The answer is really two-fold. First,