Dear Amazon,

I’m writing today to discuss the ongoing losses to Amazon shareholders, first outlined here. Unfortunately, Amazon leadership has not shown wisdom and the damage to shareholders has reached the hundreds of billions.

A Timeline and Summary of Losses

To say Amazon has not been successfully managed the past few years is an understatement.

A chart showing the stock price and business-line profitability.

Amazon shareholders are looking at over 50% losses from their peak, which constitutes a loss of all gains since 2019Q1. On the same timeline, we see the collapse of Amazon retail, both domestic and international.

I think everyone agrees this chart is a problem for Amazon and its shareholders; the rest of this letter seeks to conextualize and explain what I know of how those results came to be – based both on my own experiences and discussions with others who work(ed) at Amazon.

Driven by a Loss of Financial Discipline

For the past five years, we’ve seen increasing loss of controllership.

In 2018, Amazon’s Retail unit experienced a hiring freeze driven by profitability misses. However, few people know the story of what drove that result.

In the years proceeding that hiring freeze, Amazon’s Prime program had failed to allocate Amazon Retail their fair share of Prime’s subscription fees – due to an error in the Amazon Devices’ economic models. (Amazon Devices did the financial modeling related to customer usage of the Prime program, due to having unique customer sentiment analysis capabilities.) This math error led to routing hundreds of millions a year away from Amazon Retail into Amazon Video – approximately 30% of the Amazon Retail financial miss.

According to Jeff Wilke, the Amazon Retail hiring freeze likely could have been avoided were it not for this missing money.

In 2020, Amazon’s FinTech unit (under its Finance and Global Business Services unit) experienced severe problems. For years, Amazon leadership had failed to hire any principal software development engineers or a sufficient number of senior software development engineers to this department. This is unprofessional in the literal sense: “professions” such as medicine, law, and engineering require that junior staff be overseen by senior staff – to maintain the discipline and standards of their field.

This lack of supervision was, as of 2020, leading to routine problems: security alarms were ignored due to being triggered multiple times a day; new teams lacked senior leadership for training, so failed to meet deadlines for filings; multiple tax audits were failed, totaling tens of millions in fines, to save a few million in staffing the tax platform team with appropriate senior staff.

A lone engineer complained at this state of affairs – saying that the unprofessional environment didn’t meet the expectations of shareholders.

Amazon fired this engineer in retaliation January 2021. The engineer then wrote to the board Febuary 2021, including Jeff Bezos (chairman of the board) and Thomas Ryder (former chair of the audit committee).

Amazon’s stock price showed wobbly growth for the following five months, a moderate decline, and then a complete collapse at the ten month mark (under the uPIP program).

Because Leadership Pursued Personal Profit

One might wonder what the CFO of Amazon, Brian Olsavsky was doing while this happened.

As an employee of his throughout 2020, I can tell you: he was advancing his personal belief that we should institutionalize an endless cycle of racism. This advocacy of a return to institutional racism is often euphemistically referred to as “anti-racism” or “equity”; because its supporters (such as Brian Olsavsky) know that they’re adovcating for illegal policies – as shown by the effort of these bigots to repeal civil rights laws in California and Washington. As Ibrim X Kendi explains:

the only remedy for past discrimination is present discrimination; the only remedy for present discrimination is future discrimination

That is, the cornerstone of the philosophy Brian Olsavsky and groups such as GDIO advocate within Amazon is an endless cycle of racist retaliation and discrimination.

A return to instituional racism.

Brian Olsavsky further used his position as CFO to fundraise for the domestic violent extremists (DVEs) who brought rifles to a park in Seattle, staged an insurrection after days of violent clashes with police, built a racially segregated garden, and then gunned down teenager Antonio Mays Jr approximately a mile from Amazon HQ.

Not since the peak of the KKK have we seen Democrat-aligned executives sponsor that kind of domestestic terrorism.

And most recently, we’ve seen how that bigotry has led to Amazon’s failure on major projects:

  • The Wheel of Time
  • The Rings of Power

The failure of these properties in comparison to others isn’t explained by bigoted audiences, who accepted a strong female co-lead and black protagonists/antagonists in Reacher; but rather, by the fact that the bigotry of Amazon staff and hatred towards their own customers deeply marred those products (eg, demeaning certain demographics or erasing the myths of certain ethnicities to be “mutlicultural”). As the Dalai Lama said, “Europe is for the Europeans” – and it’s okay that mythologies, such as Lord of the Rings, retain their Euro-centric nature. Further, Wheel of Time is as a story purposefully maintains both male and female aspects to the story, and Amazon staff expressed clear misandry in removing or minimizing only the male roles.

This kind of open contempt for your own customers can permanently damage a brand in a highly competitive market.

As Part of a Broader Plan

The fashionable excuse for misappropriating shareholder resources in this manner is “environment, social, and governance” or “ESG” programs.

This intentional breach of fiduciary duty to pursue the personal political interests of corporate officers, board members, and privileged investors at the expense of common investors is under investigation by multiple states – who are also divesting from firms such as BlackRock. The role of Amazon’s leadership in this broader ideologically motivated illicit activity is essential in understanding the losses that Amazon shareholders have suffered.

One of the guiding purposes of the uPIP program is to correct these misdeeds by corporate executives.

In What Is a Stunning Act

This brings us to understanding the scale of losses attributable to those actions.

While we can’t directly quantify the losses that the CFO Brian Olsavsky caused with his involvement in the behavior noted above – and I don’t claim that it’s 100% of the losses suffered by shareholders – I do believe that it’s substantial (10%+) and that rebuilding from the losses requires stopping not only those ongoing direct losses, but replacing his defective leadership with leadership capable of maintaining professional standards. The commonality of the two financial problems described above is only two things: lax standards and the CFO.

Further, for each 1% of losses that we believe lies with Brian Olsavsky’s misconduct, we are accusing him of destroying $10 billion in shareholder assets. An Enron each percentage point.

That Will Be Challenging to Recover From

I think many of us struggle to see how Amazon will recover from these damages – which are not only the present loss of value, but damage to employee morale, brand value, etc which will impact future revenues and business.

Healing Starts With Removing the Rot

I don’t think there’s a way forward without removal of these bad actors:

From the CFO who has failed to maintain controllership while raising funds for racist DVEs, to entire organizations like GDIO and HR whose missions focus on creating the systemic racism described above, to Amazon institutions like “Affinity Groups” based on racist and sexist division, to leadership who engages in open bigotry (such as the Rings of Power staff). Looking at just the case of “Affinity Groups”, we see that Amazon has served for a generation as a bastion for racists: protecting race-based and sex-based programs at the same time Washington state was banning them as illegal discrimination – and in doing so, serving as a home for the bigots who tried to repeal those civil rights laws in 2019.

Amazon will not succeed while any of these neo-KKK members remain in positions of power:

Americans detest bigots and Amazon’s support for rebuilding institutional bigotry revulses the public.

And Then Rebuilding Relationships

Amazon has failed to maintain the relationships it needs to succeed in many regards:

  1. with customers, placing the needs of staff and suppliers ahead of customers;
  2. with employees, who aren’t supported and developed in their careers; and,
  3. with partners, who don’t support Amazon’s racism or terrorist funding.

None of these steps will be easy – and each of these will require sustained effort and investment to win back trust that was purposefully squandered.

To Return To Building the Future

Amazon has failed to lay out a strong vision for what comes next – why customers should care about what they offer. In the past, Amazon had powerful ideas for what life could be: online catalogs; global mainframes on demand. To compete going forward, Amazon is going to need new narratives developed with the stakeholders above.

And that’s what we’ve really seen these past few years: a collapse of the Amazon who could delight customers, from well-executed products to exciting visions, into a Day 2 company that focuses on fashionable biogotry and terrorism, while allowing its engineering, relationships, and products to languish. See our selected examples from Appendix B of actual ideas that Amazon leadership chose not to pursue.

Customer Obsess, Think Big, and Deliver Results

There has been a clear collapse in Amazon leadership – so I’ll end this letter reminding Amazon of its virtues; who it was before failed leaders like Brian Olsavsky.

We Obsess Over Customers

Amazon used to be defined by its obsession with customers, but has clearly strayed from that path in its Amazon Video vertical. Contrary to Japanese companies like Shonen Jump, Studio Mappa, and Studio Madhouse who focus ruthlessly on quality and customer satisfaction, Amazon (and sibling media companies, eg WarnerMedia) struggle to deliver on even established properties – due to their focus on the politics of their executives and staff.

Amazon will only succeed when it puts customers first.

We Think Big

Amazon is unable to deliver on large scale visions because politicking middle-management (and CFO) have diverted the company into focusing on fashionable bigotry rather than deliverying quality products. That incessant, small-minded nagging will never delight nor excite customers: Amazon won’t recover until these politicking groups are removed from the company to allow energy to be directed toward productive ends.

Amazon will only succeed when it removes political gadflies so it can focus on meaningful strategy.

We Deliver Results

Amazon has failed to deliver results for several years – on both strategic initiatives and routine business. That failure is due to executive misconduct, eg by the CFO Brian Olsavsky bringing the company into disrepute via funding murder a mile from Amazon HQ. I don’t believe that it’s possible for the financial leadership of Amazon to lead the company to success, given both the misappropriation of shareholder assets and the failure to maintain finacial health of the company. Brian Olsavsky has delivered negative results – and it’s time that he’s accountable for that reality.

For two years the board has stood idle while the CFO’s misdeeds wrecked the company: failure to act now is clear and actual malice, by the board.

I hope the leadership of Amazon chooses the right thing for their community and for their shrareholders – your current behavior is unacceptable. When you’re ready to make change, I’m here to help.

At Your Service,

Z Michael Gehlke

Appendix A: Partnership Chart

This chart annotates the Amazon performance chart above with our partnerships; purple represents amicable periods, while yellow denotes disputes. We show this chart to emphasize the impact we drove at Amazon – and that we initially tried to work with Amazon in private, so they could lead necessary change in the market.

Unfortunately, Amazon executives preferred bigotry to our paternship and engaged in illegal acts against us.

A chart showing the stock price and vertical profitability.

Appendix B: Roads Not Taken

I’ve included as reference a selection of ideas that Amazon leadership declined to pursure.

(2018) AWS@Home - Partnership Between Devices and AWS

Amazon Devices received a proposal to work with AWS on developing a home plug-and-play server that integrated with AWS and served as local compute for IoT or hosted services. This would serve to solve many problems related to IoT:

  • partners would no longer have to pay AWS per request for IoT, allowing for smarter devices
  • and customers would have their data stored in their home, mitigating large scale breaches
  • but developers could use familiar AWS tooling
  • and partners would pay for easily integration with backend services

This strikes a balance between a diverse needs, while allowing for AWS to integrate naturally with IoT and preparing Amazon for future technology developments. An example of such a development would be offering a Stable Diffusion application on tablets backed by this at-home compute server, to control costs: “Customers with an AWS@Home installation who use this resource for their AI engine won’t experience AWS charges per request.”

What would the future of AWS look like, if using AWS to create art during childhood was a widespread experience?

(2018) Formalized Financial Modeling

Amazon Devices received a proposal to work with Amazon’s automated reasoing group on building tools to audit financial models for sensibility at a wide scale. This proposal was in response to the failure described above, to ensure that we were testing necessary components of our mathematics and financial models. The goal was to resolve not only the immediate problem (failure to perform safety checks), but to work on resolving the problem that estimated “downstream impact” was unaligned with actual performance. The plan consisted roughly of:

  • developing a list of conditions that our model depends on
  • developing a composite model of where each financial model was used
  • ensuring that each time a report was produced, the relevant condition was checked (automated or manual)

This would pave the way to large scale analysis of Amazon Device’s finances and models; something that with five years of retrospect was badly needed.

(2020) Business Process Management / BPM-as-a-Service

Amazon FinTech received a proposal to work with our partners within Amazon Finance to build a business process tool that would better track tax filings. This was designed to address several on-going problems: difficulties applying the correct calculations; difficulties tracking all the filings in a timely manner; difficulty tracking information during audits. The idea was to build on top of AWS a platform for tracking those buisness filings and related processes (eg, processes on the filing team). A few components were proposed:

  • a graphical pipeline tracking the flow
  • which was backed by an AWS StepFunction state machine of the workflow
  • supporting subflows via invoked other state machines
  • and which was accompanied by reporting tools, eg summaries of average execution time

This would allow for a clear record end-to-end of what was done with a filing during an audit; which itself could be a second kind of state machine. The hope was that having a mechanism to coordinate between the different teams (and departments) involved with filings and audits specific to the task, rather than emails and tickets which often were cluttered or lost, would allow for better performance. (And fewer failed audits, which at the time resulted in fines of $50+ million per year.)

This would ideally be accompanied by a technical solution for each workflow component that supported better controllership.

(2020) Formalized Financial Modeling

Amazon FinTech received a proposal to work with Amazon’s automated reasoning group on building tools to audit financial models/pipelines for compliance, reliance, etc. at a wide scale. This proposal was in response to difficulties with assuring pipelines were ingesting all data and with ensuring that correct data protection rules were applied. The goal was to build a verified backend to the BPM described above – with assurance features, such as automated pipeline monitors for data quality. This consisted of three components:

  • models and verification tools for ETL pipeline frameworks, such as those used for tax payments
  • verification add-ons for financial compliance, data security, regional laws, etc.
  • automated generation of monitoring components, eg data exfiltration, loss of table integrity, or misconfigured rules dropping data

In conjunction with the previous BPM software, this would have provided a reliable basis for tax payments (which at the time, were struggling with audits).

Appendix C: What Think Big Means

I’m also including a criticism and scale-up plan related to Amazon’s upcoming projects both to emphasize how easy it is to generate better plans and to emphasize the scale of investment needed to pivot.

The Warhammer of Peace

The Warhammer property continues Amazon’s investment in top of line sci-fi and fantasy media properties. Amazon is clearly investing large sums in these properties to woo customers they’re having difficulty reaching and pleasing – yet doesn’t seem to be getting solid return on those investments. We’ve discussed above why previous efforts fell short (eg, Wheel of Time’s misandry and Rings of Power’s racism), but this time we want to emphasize that the project isn’t being ambitious enough to recover.

Amazon is facing serious headwinds and needs a multi-domain win to regain traction across silos. Further, Amazon needs a project to excite their customers both about the future and about Amazon itself. Companies that succeed on the decade scale are companies who can perceive (some of) what the world will be like in that decade and begin building for their role in that system now.

The Warhammer property presents a unique opportunity in that regard, due to its inherently multi-modal experience: building, physical gaming, digital gaming, and now media products. Amazon has a unique capability to work across all of those domains at once, from physical product manufacturing to digital gaming and media. This suggests that the best path forward would be a larger partnership.

Constructing the Future

So how can Amazon capitalize on those domains to launch a fuller Warhammer related product offering, one that can drive impact on the decade scale?

First, let’s step back and understand a little bit more about table top gaming (for those who are unaware). The Warhammer product line (both fantasy and 40k) are a series of IPs based on a table-top wargame, fought using miniature figures and dice. To accompany these products, a lore was developed – originally novels, but more recently video games and soon media products. These figures are typically manufactured by the company and purchased online or in stores, but the rise of 3D printing has led to fans creating (unlicensed) copies at home. However, not all copies are unlicensed: often, the model files are purchased and the rulebooks are purchased, but the model files used come from a different company and as such, the models are “non-regulation”. A fact that only matters in regulated tournaments at official stores.

There is a clear gap in that market for a solution around figure modeling, bespoke manufacturing, and integrations with IP. And this gap is a microcosm of the same experienced by the entire world in the move to 4th generation manufacturing. How can our manufacturing systems handle the complexity that further scientific progress and innovation will require?

Leading By Example

Amazon can create a win-win-win solution by getting ahead of this change in industry, by first doing so in the microcosm of gaming. Then Amazon will be in a position to both transition its own operstions (eg, Amazon product lines) but also to help transition the large number of small businesses that form the Amazon ecosystem.

The idea is roughly as follows:

  • Amazon deveops product line of machines
    • to develop 3D printing and small manufacturing machinery
    • both at the hobbyist (eg, garage) and shop scale
    • and controller software for these machines
    • under Amazon Devices brand
  • Amazon works with GamesWorkshop on a larger concept
    • a media show about
    • an accompanying release of a new Warhammer 40k
    • which for the first time features [GameForge] integration
    • allowing people to participate in designing units, within the universe
    • manufacturing themselves or via vendors (using above)
  • Amazon works with Titans of CNC, Mike Rowe, etc to expand program
    • introduce lessons aimed at WarHammer/toy design using [GameForge]
    • extend those lessons to integrating custom model pieces
    • extend those lessons to manufacturing at home
    • introduce clients besides [GameForge]
    • work with Titans of CNC, et al to intrgrate with other kinds of manufacturing
  • Amazon works with small businesses to simplify the design-manufacture-sell cycle
    • work with existing industry partnerships (eg, Titans of CNC shop certification program)
    • connect small businesses with them for manufacturing
    • work to simplify pipeline of requesting more units manufactured then shipping to Amazon

We can analyze this deal through a chart of the different roles involved, since there’s several participants who all benefit in different ways:

x Customers Vendors Manufacturers Society Amazon
How They Benefit Customers benefit in several ways: exciting new aspect to generations old game; increased access to modern tools; better training for modern manufacturing; better access to products on Amazon. Vendors benefit from easier access to manufacturers and tighter integrations on Amazon with the product production lifecycle. Manufacturers benefit from increased access to their customers (vendors), better tooling, and a better educational pipeline. Society benefits from increased manufacturing education and increased access to modern manufacturing. In particular, teaching children through games is a long term investment. Amazon benefits not only in the direct product lines (devices + Warhammer IP), but through flywheel effects in increased vendor offerings and product diversity.
What They Provide Demand for products and requests for bespoke manufacturing. Products for sale and requests for manufacturing. Manufacturing products per request. – Manufacturing equipment and transactional framework.

People expect this kind of pro-social dealmaking and future crafting from their largest businesses.