Product Research Data Accuracy Revisited


Joe Junfolla of SixLeaf is with us explaining his view on why data accuracy matters for Sellers,, his experiences as a seller with inaccurate data, the error rates from some of the tools out there and what SixLeaf are building as an algorithm to help.

  • Anthony Lee, good friend of the show, works with Joe at SixLeaf
  • Joe is CEO/Founder of what was ZonBlast
  • Seller since 2014
  • Pivot in 2017 now to SixLeaf
  • Specialise in data accuracy and ensuring the end user (Seller) gets info to make accurate decisions in their business

Facebook Group

  • #Brand Owners by Six Leaf
  • Used to be Zon Squad as a paid group
  • Opened it up as a good close knit group
  • No Gurus allowed 🙂

Jungle Scout data

  • Had Greg Mercer on the show on Jungle Scout data accuracy
  • Gave out to people a load of data
  • The SixLeaf group has a view on the data provided

Respect for the Space

  • Lots of respect for Jungle Scout and Viral Launch
  • Going back to 2014 there was a sales estimation tool being used for a training course that was incredibly inaccurate on sales per day/month (was 4-6 times over the actual sales)
  • Underlying theme is that these chrome extensions are based on estimated sales
  • Discussions in groups over the years have revolved around numbers not being accurate to them being used as a ‘relative measure’ between ASINs
  • Got under our skin as the numbers are not accurate but Noobs would follow them slavishly

Solving the Problem in the Space

  • 2016/17 the chrome extensions space was blowing up
  • Six Leaf wanted to get into this space but was held back by the estimated sales figures as didn’t think could get better than what was out there
  • Also, wanted to focus on being unique and not a me-too
  • 8 months ago learned about machine learning and creating an accurate sales algorithm
  • Dove in and in December contracted with data scientists that specialise in Amazon
  • Literally 100 of millions of rows of data to create this learning algorithm for BSR, search volume nad history for a given ASIN
  • Wanted a 15% target error rate and seem to be hitting 6% which is amazing
  • Incredibly precise and accurate

Accuracy versus Precision

  • The challenge of data accuracy in these numbers will be great
  • Difference between accuracy and precision
  • Think of archery and arrows
  • Accuracy is how close you are to the bulls-eye
  • Precision is how tight your shot-group is
  • Basically – everybody is very inaccurate and very imprecise
  • Box-plot with whiskers shows this argument and that Six Leaf are the top at the moment

Phoenix Algorithm

  • Traditionally we’ve been niched in launch
  • Moving more into SaaS now with the new Phoenix Rise algorithm

Launching Soon

  • Data scientists and Greg is spending $1m per year to develop data accuracy
  • Joe’s approach is different here
  • Throwing money at something does not definitely make it better
  • Joe not spending anywhere near $1m on this
  • Has a beta group and internal team running this, around $10k/month
  • Launching open Beta in the next week

Contact

  • #BrandOwners group on Facebook

Make your VOTE count!

Your community needs you!

We will be in touch shortly with further details. Don't for get to share this page with the community.

Share This
%d bloggers like this: