0
0
Back in 2012, long before “AI” was in every pitch deck, I founded my first AI startup, Nyopoly.com.
The history of retail commerce is a history of pricing mechanisms. From the haggling of ancient bazaars to the fixed-price revolution of the 19th century, and finally to the dynamic algorithmic pricing of the digital age, the method by which value is assigned to goods has constantly evolved. In 2012, I found myself at the intersection of these eras when I founded Nyopoly. With deep roots in both legacy retail (TJX) and technology (IBM), I sought to operationalize the theoretical “holy grail” of microeconomics: Perfect Price Discrimination.
This report reflects on my journey, from the genesis of the idea to the technical architecture we built, and the ultimate legacy of the company. It reconstructs the trajectory of a startup that attempted to replace the adversarial nature of auctions with a “non-competitive,” learning-algorithm-driven negotiation engine.
Looking back, I can see that Nyopoly was not merely a niche fashion startup but a precursor to the modern era of “Customer Engaged Pricing” (CEP). I identified a $2 trillion inefficiency in the global retail market—the “deadweight loss” of fixed pricing, and attempted to capture it. This retrospective culminates in the three critical lessons I learned regarding the cognitive friction of negotiation, the strategic reality of needing B2C proof for a B2B product, and the “cold start” data challenges of early AI.
Part I: The Genesis of Disruption – The Retail Context of 2012
1.1 The Post-Crisis Consumer Psyche
To fully appreciate the radical nature of what we were trying to build, you have to remember the psychological and economic reality of the American consumer in 2012. The Great Recession had fundamentally altered the DNA of shopping behavior. The era of conspicuous consumption had given way to the era of the “smart shopper.” Value was no longer just about the product; it was about the victory of securing a deal.
Between 2009 and 2011, this behavioral shift fueled the meteoric rise of the “Flash Sale” model. Companies like Gilt Groupe and Rue La La exploded onto the scene. They gamified shopping, creating an adrenaline rush associated with “winning” a limited-inventory item.
However, by 2012, I saw that the flash sale model was facing an existential crisis. Consumers were experiencing “deal fatigue,” and luxury brands were becoming wary that deep public discounts were eroding their brand equity. The industry needed a “safety valve” for excess inventory that did not publicly devalue the brand.
1.2 My Background
I didn’t view this purely as a software problem; I viewed it through the lens of a seasoned merchant. My background positioned me to see the friction between rigid pricing and inventory liquidity clearly:
- ECC Life & Style: As co-founder and CEO of this luxury menswear company, I experienced the pain of the seller firsthand: set the price too high and sit on inventory, or set it too low and leave margin on the table.
- TJX Companies: Working with the parent company of T.J. Maxx and Marshalls gave me a masterclass in “opportunistic buying”, moving massive volumes of inventory that full-price retailers couldn’t sell.
- e4eNet: My first significant startup where I served as VP of Operations developing a first of a kind “cloud” platform for the collaboration on the Design for manufacturing process for Original Equipment Manufacturers (OEM) and Contract Manufacturers (CM), sold to IBM.
- IBM: My time here provided the technological rigor, specifically in “Smarter Commerce” and supply chain optimization.
My academic foundation, a BS from UMass and an MBA from Babson College, helped me synthesize these experiences. In 2013, my selection as a Fellow in the Startup Leadership Program in Boston further signaled that we were onto something significant.
1.3 The Core Thesis: The $2 Trillion Inefficiency
The intellectual bedrock of Nyopoly was a staggering statistic I couldn’t ignore: 40% of all retail industry revenue is driven by discounted transactions, resulting in $2 Trillion in lost revenue.
I realized that the “single-price” model was broken. In a physical store, you cannot change the price tag for every person who walks in the door. But online, the price tag is just pixels. It can be dynamic.
My mission was to build a “pricing engine” that could capture both the customer willing to pay a premium and the bargain hunter. By allowing one customer to pay $150 and another to pay $90, we could theoretically capture the entire area under the demand curve. This was the birth of Customer Engaged Pricing (CEP).
Part II: The Economic Engine – Deconstructing Perfect Price Discrimination
2.1 The Hierarchy of Price Discrimination
To understand why our “learning algorithm” was so ambitious, we have to look at the microeconomic theory of Price Discrimination.
- Third-Degree (Segmentation): Discounts for students or seniors. It’s blunt and inefficient.
- Second-Degree (Self-Selection): Bulk discounts or auctions. In an eBay auction, the price is determined by the second-highest bidder, meaning the winner still pays less than their absolute maximum.
- First-Degree (Perfect): This was our goal. Charging every single customer exactly their maximum Willingness To Pay (WTP). If we could achieve this, we would maximize profit and eliminate “Deadweight Loss.”
2.2 Our “Non-Competitive” Innovation
A critical differentiator for Nyopoly was our rejection of the auction model. I explicitly positioned it as: “It’s not an auction, there is no commitment, and you’re not competing against other members”.
Why was this distinction vital?
- Psychological Safety: We wanted to remove the fear of the “Sniper” and the stress of competition. It was a private negotiation between the user and our algorithm.
- Inventory Depth: Auctions fail for depth inventory (500 units of the same bag). Our model allowed us to sell identical items to different people at different prices simultaneously.
- Brand Protection: This was the “Killer App” for the B2B side. Luxury brands hate auctions because the closing price is public. On Nyopoly, the price was a private contract.
2.3 The “Engaged Pricing” Workflow
We designed the user flow to be a game of strategy rather than a passive transaction:
- Discovery: The user browses our curated selection.
- Valuation: Instead of a price tag, the user answers: “What is this worth to you?”
- The Offer: The user inputs a price.
- The Algorithmic Decision: Our engine processes the bid against the floor price, demand curve, and user probability.
- The Response: Either “Congratulations!” or a nudge to try again.
We transformed the passive act of spending money into an active, “customer-engaged” event.
Part III: The Technological Architecture – The “Learning Algorithm”
3.1 The State of AI in 2012
We were building this on the eve of the Deep Learning explosion. The “learning algorithm” I referenced wasn’t the generative AI of today. It was built on Reinforcement Learning (RL) principles.
3.2 How I Built the Algorithm
Our engine functioned on a “Multi-Armed Bandit” logic:
- Objective: Maximize Profit over time.
- State Space: We looked at product attributes (seasonality, inventory) and user attributes (referral source, past bidding history).
- Action Space: Accept, Reject, or Counter-Offer.
The core challenge was balancing Exploration vs. Exploitation. If a user bid $100, was that their max? Or would they have paid $150? If we accepted immediately, we left money on the table (exploitation failure). If we rejected it to push for more, we risked them leaving (exploration risk). Our system “learned” by tracking the success rates of these rejections to find the optimal threshold.
3.3 The “Cold Start” Problem
The biggest hurdle was the “Cold Start” problem. To discriminate effectively, the algorithm needed data. For a new user, we had zero history. We had to rely on proxy data (zip codes, device types, time of day) to make initial guesses, essentially moving from perfect 1st-degree discrimination back toward segmentation until we learned more about the individual.
Part IV: The Pivot – Proving the B2B Engine through B2C
4.1 The Strategic Reality
While Nyopoly is often remembered as a consumer site, my original vision was always B2B. I wanted to sell the “Engaged Pricing” engine to major retailers.
However, I quickly ran into a wall: big retailers are risk-averse. They weren’t going to plug a “black box” pricing engine into their multi-million dollar ecommerce stacks without proof. They needed to see it work in the wild.
4.2 Building the “Lab”: Nyopoly.com and GetWhatYouLove.com
To prove the thesis, we had to build the store ourselves. We launched Nyopoly.com and subsequently acquired GetWhatYouLove.com to complement it. These weren’t just shops; they were live laboratories for our technology.
We used a “members-only” gate for tactical reasons:
- Exclusivity: It mimicked the “velvet rope” of Gilt Groupe.
- SEO Cloaking: By hiding prices, we prevented Google from indexing our low negotiated rates, allowing us to sell current-season merchandise without violating Minimum Advertised Price (MAP) agreements.
4.3 Validation: The New York Fashion Tech Lab (2015)
The strategy worked. The data we generated from our B2C operations validated the engine, leading to our acceptance into the New York Fashion Tech Lab (NYFT Lab) in 2015.
We weren’t just a scrappy startup anymore; we were connecting with the establishment. Our cohort included data analytics firms like 42 Technologies and social merchandising tools like InSparq. The Lab explicitly positioned us as “Price optimization at the customer level… unbiased and non-competitive”.
Presenting at our Demo Day at Time Inc. in June 2015 was the culmination of this strategy. Pitching to executives from Estée Lauder and Ralph Lauren, I finally had the data to back up the B2B pitch: “Engaged Pricing” was the antidote to the race-to-the-bottom.
Part V: Three Lessons Learned from the Trenches
Reflecting on the rise and evolution of Nyopoly, three specific lessons stand out. These are the things I wish I had known in 2012.
Lesson 1: The Physics of Human Change, Should Not Be Underestimated.
The Insight: Economic models assume consumers are “Rational Utility Maximizers.” In reality, they are often “Cognitive Misers” who will pay a premium to avoid thinking.
My Experience:
My value proposition was “Negotiate your own price.” To an economist, this is empowering. To a busy mother of two, this is work. The standard fixed-price model is frictionless. Our model introduced a transaction cost: “Am I bidding too high? Will I be rejected?”
Even if we saved the user money, the mental effort reduced the “net utility” of the transaction. I learned that for many, the certainty of a fixed price is a feature, not a bug.
Lesson 2: Good Data Matters.
The Insight: AI is only as smart as the volume of data it consumes.
My Experience:
Our algorithm was designed to predict WTP. When it worked, it was magic. We observed that, on average, consumers were willing to pay 20% more than what they would in a standard markdown plan. This was the “found money” I had predicted.
However, achieving this consistently required massive scale. Without the data velocity of an Amazon, our algorithm faced high variance. It was difficult to make granular pricing decisions for every new user immediately. We were building the Ferrari engine before we had the fuel line fully laid.
Lesson 3: Micro Pivots are a Part of the Journey.
The Insight: You cannot easily be a disruption to the industry and a vendor for the industry simultaneously without a clear strategy.
My Experience:
We started as a B2B play but had to bolt on a B2C destination to prove it. This created a complex identity.
- As Nyopoly.com: I was competing with Macy’s.
- As Engaged Pricing: I was trying to sell software to Macy’s.While the B2C site was necessary for validation, it created channel conflict. Retailers are paranoid about data. Successful pivots require extreme focus. I learned that sometimes you have to build the store to sell the technology, but you must be ready to drop the store the moment the technology takes off.
Part VI: The Legacy
Nyopoly was an ambitious experiment in behavioral economics. We challenged the 150-year-old hegemony of the fixed price tag and correctly identified that the internet should enable a more fluid market.
Today, we see the evolution of my idea in “Surveillance Pricing”, where retailers implicitly adjust prices based on user data. But unlike the covert methods used today, my vision was transparent: “Customer Engaged Pricing.”
My journey from retail executive to tech founder has been defined by this pursuit of value and efficiency. Whether optimizing a supply chain or now project management with Coworked Harmony, the principle remains the same: systems should work with the people who use them.
