superfood picker

Your New Best Friend: The Superfood Picker – Algorithmically Compare Greens Powders

superfood pickerDudes, dudettes………allow me to humbly present to you the fruits of a nerdy fetish I’ve been incubating for 6 months:  The Superfood Picker.

It’s an interactive javascript program that allows you to compare greens powders side-by-side and algorithmically rate them based on your preferences.

Its announcement is extremely gratifying because its production has been a black hole where fun goes to die in my personal life these last 183 days.  I keep putting off social engagements because each weekend I find myself staring at my computer screen, with brows sweaty and tongue parched, determined to patch the memory leak that tacks on 0.00018 seconds to the page load time.

Women are beautiful, but dammit, if I don’t organize the world’s information on greens powders then who will?  It was my occupational providence to get ‘er done.

So now that it’s finally here I figure this oughta’ appease their doubts since they can see the sheer importance of my intellectual rabbit holes.

“You see guys?!  Now you can totally grade esoteric health supplements with the click of a button!  THE WORLD HAS CHANGED.”  

…….Or maybe not.  Anyways……

It was hand coded by yours truly with the goal of creating an easy way to compare superfood products interactively using objective information that can be easily verified.  It also acts as a portal for each product that allows you to explore what the web has to offer about each one.

How It Works

If you want you can simply compare any two products you’d like side-by-side and make an apples-to-apples comparison about their price, types of ingredients,etc by going to the middle of the page:

superfood picker

Each product also has its own dynamically generated product page that aggregates all of its relevant information that you can get to by clicking on its block:

superfood picker 2

However, to me the most interesting part of the program is the analytical engine that rates superfood powders based on the preferences you give it.

Let’s take a minute to discuss how it works.

Three Scores:  Price, Taste, Nutrients

The superfood picker generates three different scores for each product in its database from user input:  a price score, taste score, and nutrient score.  Each one is from 1-100, but they’re weighted according to your preferences in the following way:

  • 1st preference:  50% of the score.
  • 2nd preference:  30% of the score.
  • 3rd preference: 20% of the score.

So if a product has a taste score of 80, nutrient score of 90, and price score of 70 and your priorities are taste, price, nutrients in that order its final score will come out as follows:

  • 80 * 0.5 = 40 (weighted score of taste)
  • 70 * 0.3 = 21 (weighted score of price)
  • 90 * 0.2 = 18 (weighted score of nutrients)

With a total score of 40 + 21 + 18 = 79.

Here’s how each score is determined:

Price Score:

This one is simple.  It compares your desired price with the product’s price and penalizes a product in a linear way depending on the difference between the two.  The penalties are more steeply levied depending on your priority.  (Ie, someone who states price as  their least important priority will not have the price scored penalized as much as someone who lists price as their biggest priority if there’s a discrepancy between the two).

Taste Score:  

Each product was evaluated for its sweetness, its amount of a “greens” taste, and whether or not it has flavors added to it.

These characteristics are then matched to the users stated preferences.  “Don’t Care” is treated the same as being in 100% agreement with whatever characteristics the product has.   There’s a bit of subjectivity here, since what constitutes “sweet” will vary from person to person.  To compensate the algorithm treats small differences between user preferences and product characteristics as negligible and the penalties get steeper the bigger the differences are.

So if a user has a preference for something that’s “modestly sweet” and a product is graded as being “pretty sweet” it assumes the match is pretty close and that product would still get a high taste score.  But if someone has a preference for “as sweet as possible” and a product is graded as “not sweet at all” it assumes this difference is easier to discern and punishes the product more severely.

Nutrient Score

The nutrient score is the most complicated because it has two dimensions:  the innate nutritiousness of a product and the desired ingredients or characteristics a user wants.

So the superfood picker makes two nutrient scores:  an internal one based on a product’s characteristics and external one that measures how closely it matches user preferences.

The internal score evaluates a product’s ingredients and features and creates its own interpretation of how nutritious the product is.  This is important because most people don’t know exactly what they want and need the superfood picker to do some thinking for them.  The match score is a measure of how closely its ingredients or features match what you want.  (It’ll also disregard your preferences if you don’t have any).

The Internal Nutrient Score

The internal nutrient score has three components:

  • It gives a bonus if a product is organic or wildcrafted.
  • It penalizes a product if it has a disproportionately large portion of its ingredient list comprised of low quality ingredients.
  • It rewards products that have a wide variety of nutritious ingredients.

Every product has its own internal nutrient score regardless of the preferences you enter into the program.

Notice it doesn’t give an explicit preference for the relative amounts of different nutritious ingredients, or for being vegan, paleo, kosher, etc.  It assumes that if people specifically want these they’ll make note of them in their preferences and thus improve the external nutrient score for products that have these qualities.

If you don’t state any specific preferences the final nutrient score will be the internal nutrient score.  Otherwise it’s a blend of the internal and external scores.

The Superfood Picker Has Strengths And Weaknesses

Like all computer programs the superfood picker has its blind spots.

Here’s my opinion on what it gets right and wrong:

  • It keys in on the most important aspects of a greens powder.  I make a greens powder and have had tens of thousands of people come to my website to find information on how they work.  Over the years this has created a good intuition about what people ought to look for in them and it does a good job of keying in on these qualities even if the user doesn’t know what they want.  It also gives discerning customers the ability to dissect products in a very particular way if they have specific tastes.
  • It underrates products that are sold in large bulk quantities.   There are some very good products that have a high sticker price but very reasonable per ounce price.  These products are underrated by the superfood picker because it grades products based on  their 30-Day-Supply price, regardless of portion size.  That’s because most people know how much they’ll pay for a 30 day supply of something but have no idea what a good per ounce price is.   You can compare products based on their per-ounce price but the algorithm doesn’t grade products on that criteria.
  • It underrates certain niche products.  The superfood picker is calibrated to the “middle” of the greens powder curve, which means it grades products against assumptions that there are certain classes of ingredients a greens powder ought to have.  This means some super-specific nutrient powders that disproportionately emphasize one ingredient or specific classes of ingredients might have their nutrient scores punished in a way that’s not entirely accurate.

Despite these flaws I think the program does a good job of getting the basic idea right.  If you answer the questions honestly you’ll more than likely get a recommendation that’s on the mark.

It also brings a lot of transparency to this corner of the supplement world.  This brings me great satisfaction because it’s difficult to compare products due to their long list of ingredients and different health claims.

But now the world has an objective way to peek behind the curtain and evaluate the landscape in a way that’s suited for them.

Other Important Stuff

Relevant info you ought to know about:

  • My products are not listed in the program.  This is admittedly a controversial point but I was worried about bias (or accusations of bias) creeping in and casting doubt about the intellectual integrity used to make the program.  Without my products listed it’s easier for the program to have a level playing field that performs its most important function well:  helping people evaluate all the other greens powders out there with accuracy.  That said, I’ve run Incredible Greens  through it and it tests out most closely to Macro Greens.  They have similar ingredient bases and flavor profiles so it’s safest to assume what’s true of it would also be true of Incredible Greens.
  • It generates affiliate links.  One motivation for the program was to create a revenue channel for Health Kismet outside of physical product sales.  To do this some of the links to Amazon are affiliate links and I’ll get a 4% commission when you click on them and buy something.  Hopefully that doesn’t bother you.   One of the challenges in building my company is finding ways generate enough profits to continue to fund new projects.  It’s hard to do this through product sales alone because I make about $10-$14 an order when all is said and done.    The most common way get additional money is to take out a loan or find investors.  At the moment I don’t think I could do this without diluting what makes my company unique so I have to find other places to create additional revenue.  The superfood picker is one such attempt.

You can find it in the “Tools” section of the navigation bar like so:

superfood picker 3

Please do go check it out and let me know what you think.

That is all!

1 thought on “Your New Best Friend: The Superfood Picker – Algorithmically Compare Greens Powders”

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s