Weekly reading – 10th April 2021

What I wrote last week

Get back to what you love

My experience so far with Amazon Shopper Panel


An interview with an Apple veteran who shed some light on the culture of secrecy

Supreme Court Sides With Google In Decade-Long Fight Over API Copyright; Google’s Copying Of Java API Is Fair Use. If you have time, look for the opinion written by Justice Breyer on API. It’s good!

Amazon Global Supply Chain and Fulfillment Center Network. Just look at the number of fulfillment centers and warehouses Amazon has!

How We Bootstrapped a $1M ARR Email Client

Shopify: The E-commerce On-Ram‪p‬. I may have found another favorite podcast. The first episode on Shopify doesn’t disappoint!

9.5 million customers traded cryptocurrencies on Robinhood in Q1 2021, compared to 1.7 million in Q4 2020

What I found interesting

A landmark study showed promising results that could help us produce a vaccine for HIV. A remarkable time to be alive. You gotta admire the work that scientists around the world put in.

Barrier Reef doomed as up to 99% of coral at risk, report finds. “The Great Barrier Reef is all but doomed, with between 70 and 99 per cent of corals set for destruction unless immediate “transformative action” is taken to reverse global warming, according to a new report. The Australian Academy of Science says the more ambitious target of the Paris Climate Agreement of keeping global warming to 1.5 degrees has now slipped out of reach and is “virtually impossible”.

What the U.S. Can Learn From China’s Infatuation With Infrastructure.

Apple has been granted a patent for Systems relating to a National ID Verification System

NYTimes’ profile of Katalin Kariko, the scientist whose work on mRNA helped save the world from Covid-19

Interesting stats

Lithium battery costs have fallen by 98% in three decades. If I have a kid, he or she will likely have electric vehicles as cheap as ones with a combustible engine.

7% of America’s population or more than 2.3 million Americans don’t use the Internet

Amazon reached 10% of the US advertising market

Plant-based food market grew to $7 billion in 2020, up 27% year over year

80% of Europe’s in-store transactions are now contactless, according to Mastercard

E-commerce Evolution in the US by Mastercard
Source: Mastercard

My experience with Amazon Shopper Panel

I wrote about Amazon Shopper Panel before. The program is on an invite-only basis. Essentially, participants upload 10 non-Amazon receipts (Whole Foods transactions aren’t counted either) every month to earn $10 in Amazon balance and have an opportunity to earn more by completing surveys. In my post, I wrote about the immense applications that can come from this initiative. This time, I want to update you with more details on the program. I myself received an invitation back in February 2021. If you are selected, you will receive an email like this

Amazon Shopper Panel Invitation
Figure 1 – Amazon Shopper Panel Invitation

The app is fairly simple. The first tab gives the user an overview of how much in rewards he or she has earned so far every month. The second tab is where receipts can be uploaded while the third tab houses all the surveys that Amazon wants you to complete. There are other routine sections such as FAQ, Contact Us, Legal Information and Sign Out that are tucked in a window that will open once you touch the three-dot symbol.

How Amazon Shopper Panel App Looks
Figure 2 – How Amazon Shopper Panel App Looks

Receipts can be uploaded via a phone camera. Based on my experience so far, the app is fairly receptive towards even wrinkled receipts and those that have small tears. Email receipts are qualified, as long as they are sent to receipts@panel.amazon.com from the same email that a participant uses to register an account. I got a car wash voucher from a dealership a while back and used it at a random fuel station in Omaha. The receipt from the car wash was still accepted, much to my surprise, because it didn’t have any card information. Since February 2021, I completed two surveys and received 25 cents for each. The surveys featured only one question each time and it was pretty basic such as, I paraphrase here, “where did you get information for your online purchases?”.

 Email receipts are accepted
Figure 3 – Email receipts are accepted
Even a car wash receipt without payment is accepted
Figure 4 – Even a car wash receipt without payment is accepted

Since starting to use the app, I have paid attention to how receipts differ from one another in terms of structure and layout. The computational process used to digest these receipt images will have to be pretty sophisticated to handle the intricacies and variety in how receipts are printed and captured. If Amazon can gain this ability to read images, it can be applied to other parts of their retail business.

Even if receipts are input by humans, the intelligence that Amazon may gain from this initiative will open up a lot of opportunities:

  • Design new private label products
  • Court other retail partners with unprecedented and reliable data
  • Support their ads business
  • Upsell current customers by understanding them better

My expectation is that Amazon will select enough folks from different backgrounds to join this effort. The participants have to be representative of consumers in America in terms of age, race, income and gender. Plus, the pool has to be big enough and the time period should be long enough so that the data can be statistically significant. As a result, at $10/participant/month, this initiative can reach 6 figures pretty fast. If the time and resources dedicated to the analysis task are factored, the expense will rise even higher. To other smaller retailers, the technical and financial barriers are not easy to overcome. To other retail giants like Walmart, I am surprised not to see a similar effort from them. This is the type of initiative where if your rival gains the first mover advantage, it will be a tall order to claw it back.

In life, there are skillsets which are very difficult to gain, but once mastered, can offer long-term leverage in various aspects of our life. Think: sales, writing, coding, human languages, cooking, fitness. In my opinion, this initiative belongs to the same category. If it proves to be successful, Amazon Shopper Panel can arm Amazon with intelligence and capabilities that are going to lift the company to even greater heights.

Disclosure: I have a position on Amazon and Walmart in my personal portfolio.

Weekly readings – 24th October 2020

What I wrote last week

How Apple Card’s balance grew by 50% in 3 months and the implications

Amazon Shopper Panel


EU is shooting for an ambitious “industrial cloud” plan to rival the US

DOJ filed an antitrust lawsuit against Google and the tech giant had a strong rebuke

Will you exchange privacy for some money every month? If yes, Amazon will pay you $10/month for 10 receipts of non-Amazon purchases through a service called Amazon Shopper Panel

Doordash is still the market leader in the meal delivery space with 49% market share in September 2020, according to SecondMeasure

A fascinating story on a guy who learned, worked and blogged his way to be an authority in the podcast space

DOJ’s lawsuit against Google, a very interesting read with a lot of great information

Boosted by better-than-expected consumer spending and write-offs, card companies are eyeing more customer acquisitions

How Apple is organized for success


Emerging Architectures for Modern Data Infrastructure

Adobe released a new feature to help creative folks get credit for their honest work and fight misinformation

a16z released an interesting blog post on the promise of Payroll API

What I found interesting

Japanese Butter Tableware. Ain’t they beautiful and interesting?

How Egypt is growing forests in middle of the desert

A damning account of the failed project between Foxconn, Trump and Wisconsin. The red flags have been there for a long time, yet I fear this isn’t the last time we hear something about it

Amazon Shopper Panel – Amazon’s latest big bold move

Yesterday, Amazon announced a new initiative called Amazon Shopper Panel, which I think can have major ramifications for the company moving forward. Amazon Shopper Panel (ASP) is an opt-in and invite-only program in which participants can earn rewards by sending at least 10 eligible receipts a month to Amazon. To be eligible, receipts have to come from non-Amazon purchases, excluding even transactions at Whole Foods, Amazon Books or Amazon Go. There is no language on the minimum value that a receipt has to reach to be eligible. If a participant successfully sends 10 receipts to Amazon in a month, he or she can earn $10 in Amazon Balance or as a donation to a charity organization of their choosing. ASP’s participants can opt out at any time and delete previously posted receipts. In addition, participants can earn more rewards by completing surveys. There are few disclosures on what the surveys will be about and how much responders will earn for each survey. But my guess is that they will be aiding Amazon to complete a shopper’s profile with demographic information.

In my opinion, this is a consequential move for Amazon for the following reasons

Build an unmatched shopper database

As a retailer, one of the biggest challenges is to understand who your customers are and what they do outside of what transpires between you and them. If you are a Target shopper holding a Target-branded credit card, all they know about you is what takes place on Target’s website and between Target’s store walls. With some effort, Target may work with their credit card issuer to understand a bit about your spending behavior on your credit card. Unfortunately for Target, they have no idea on other parts of your consumer life and what you do. And you can bet that they want to know!

By compiling and analyzing receipts on a regular basis, Amazon can build a powerful and formidable database. Receipts carry a lot of information. Not only do they tell where you shop, when, how much you pay, how you pay and the frequency, but receipts also reveal what you buy down to an item level. Essentially, Amazon can theoretically know what vegetables, fruits or milk you buy every week. Normally, retailers can at best gather data down to a merchant level such as Walmart or Target. However, I don’t know of any retailer in the US that can basically gather intelligence down to that level of details.

Enable more sale on Amazon.com and Amazon’s private labels

The first application would obviously be to increase sale on Amazon.com, specifically the sale of Amazon’s higher-margin private labels. The intelligence Amazon gathers from ASP can help them decide which products can be produced as private labels, which can be marketed to whom and at what time. Information is worth as much, if not more, than money. Amazon operates at a scale that even a good idea or a degree of increased efficiency can mean millions of dollars.

Their Amazon Basics brand now has 3,000 SKUs.

Improve their ads business

Though small as a component of Amazon’s revenue, advertising is now a business with $20+ billion annual run rate and the latest YoY growth of 41% (as of Q2 FY2020). Advertisers like ads on Amazon because Amazon shoppers already have intention to buy, something that is lacking on Facebook and Google. When you search for something on Google or scroll down your Facebook Newsfeed, the intention to buy isn’t always there. However, if you go to Amazon, it’s likely that you will end up with an order and that’s what advertisers want. Though impressions are great, advertisers prefer much more hard cash coming from new purchases.

ASP gives the strongest and most reliable signal as to what shoppers are willing to pay for down to the item level, how much and when. Sophisticated and powerful as Facebook and Google are, they don’t have a system in place to know what shoppers buy in store. On the other hand, Amazon would be in a unique position to help advertisers place an ads to the right audience with the right message and at the right time. And to charge a little premium on that service as well. An expansion in advertising means an expansion for both revenue and gross margin for Amazon as this segment is surely more profitable than their online store.


The understanding of consumer behavior in physical stores carries a lot of value and offers potentially great applications. I am confident that somewhere some smart heads at Amazon already came up with more use cases than what I laid out above. Moreover, this kind of capability is pretty hard to emulate. To gain the intelligence that Amazon hopes to achieve, a company needs to have financial and technical resources that Amazon possesses. Having photos of receipts is one thing. Parsing out information and turning that into actionable intelligence is a completely different matter. while this move may raise red flags for privacy concerns, it’s a shrewd business move. In my opinion, there’s no better way to create a moat in retail than building up a deep understanding of consumers and leveraging your size, technical and financial advantages. I, for one, look forward to what this move will bring about.

Disclosure: I own Amazon stocks in my portfolio.