TL;DR for sellers
- Jeff Bezos says most product ideas never ship because turning an idea into a real product means hiring people, forming a company and carrying huge upfront costs. It's the same kind of barrier Amazon has spent decades knocking down for sellers.
- For e-commerce sellers, the gap is already smaller than it's ever been. The real killer isn't the cost of building. It's paying to build something before you know anyone wants it.
- The fix is a staged testing ladder: desk research, complaint mining, smoke tests, content tests, pre-orders, micro-batches, then a scale order. Each rung costs more, so you only climb when the last one passed.
- Set your kill, iterate and scale numbers before you start a test, not after. Otherwise every result looks "promising."
- Most good tests cost under $500 and take under 30 days. Most expensive flops skipped them.
The quote that every seller should take personally
In a recent Fox interview, Jeff Bezos put his finger on why so many product ideas never make it past the "wouldn't it be cool if" stage:
"Probably everybody watching at some point had an idea for a product. The reason you didn't build it is because it's too hard...you have to hire people and form a company."
His point is that imagination isn't the bottleneck. Infrastructure is. And he wants to make the distance between an idea and actual manufacturing much shorter.
If that sounds familiar, it should. Shrinking the distance between an idea and a customer is basically the Amazon playbook. Marketplace let anyone sell to millions of shoppers without opening a store. Fulfillment by Amazon meant you didn't need your own warehouse or shipping team. AWS meant startups didn't need to buy their own servers. Each one took something that used to require "hiring people and forming a company" and turned it into a service you could switch on.
Bezos also built Amazon around cheap experiments. He's long said that if you double the number of experiments you run, you double your inventiveness, and that failure is part of invention. Plenty of Amazon ideas flopped (the Fire Phone being the famous one), but the company could afford to fail because each test was small relative to the wins.
But the principle behind it is exactly what separates sellers who launch winners from sellers with a garage full of unsold stock: make testing an idea so cheap and fast that you can afford to be wrong a lot.
The seller version of Bezos's problem
For a private label or DTC seller, "too hard" rarely means hiring engineers. It usually looks like this:
- A factory MOQ of 500 to 1,000 units before you've sold one
- Thousands tied up in inventory, freight and duties
- Weeks of photography, listings and ad setup
- A launch that flops, followed by months of discounting to claw back cash
Bezos's insight, reframed for sellers: the cost of testing an idea is too high before you've proven it's worth anything. So most sellers either never test (and never launch), or skip testing (and launch the wrong thing).
The good news is that the tools to test cheaply already exist. Landing page builders, short-form video, print-on-demand, low-MOQ suppliers, pre-order apps and marketplace listings let you get real buyer signal for a few hundred dollars. You just need a system.
The Idea-to-Shelf Testing Ladder 🪜
Think of product testing as a ladder. Every rung costs more money and time than the one below it, and every rung gives you a stronger signal. The rule is simple: you only climb when the rung you're on gives you a clear pass.
| Rung | Test | Typical cost | Time | What it proves |
|---|---|---|---|---|
| 1 | Demand desk research | $0 to $50 | 1 to 2 days | People are already searching and buying in this space |
| 2 | Complaint mining | $0 | 1 to 2 days | There's a real, repeated gap your idea fixes |
| 3 | Smoke test (landing page + ads) | $100 to $400 | 5 to 7 days | Strangers care enough to click and sign up |
| 4 | Short-form content test | $0 to $200 | 7 to 14 days | The idea is interesting enough to stop a scroll |
| 5 | Pre-orders or paid waitlist | $50 to $300 | 7 to 21 days | People will actually pay, at your price |
| 6 | Samples + micro-batch (10 to 100 units) | $300 to $2,000 | 2 to 6 weeks | The product can be made to spec and buyers like it in hand |
| 7 | Marketplace micro-launch | $200 to $1,000 | 2 to 4 weeks | It sells to cold traffic with real margins and reviews |
| 8 | Scale order | $3,000+ | 6 to 12 weeks | You've earned the right to commit real cash |
Costs and timings are ShelfTrend working estimates for a typical small private label seller. Your category, platform and region will shift them.
Let's walk through each rung.
Rung 1: Demand desk research
Before you spend anything, check that the market exists. You're not looking for proof your exact product will sell. You're looking for proof that people already spend money on the problem.
Tactics:
- Search volume and trend direction. Google Trends, marketplace autocomplete and keyword tools tell you if interest is rising, flat or fading. A rising five-year line beats a big but shrinking one.
- Sold listings, not active ones. On eBay, filter to sold items. On Amazon, look at Best Seller Rank and review velocity on similar products. Lots of listings with few sales is a warning sign.
- Price band check. Note where the bulk of sales sit on price. If everything sells under $20 and your idea needs to retail at $60, you've found your first big risk.
- Social proof of interest. Search TikTok, Instagram and Pinterest for the problem your product solves. Are people posting about it, saving it, asking where to buy?
Pass if: you can find steady or rising demand, multiple sellers making real sales, and a price band that fits your likely costs.
Rung 2: Complaint mining
This is where most winning private label ideas come from, and it's free. Buyers tell you exactly what's broken. Your job is to listen at scale.
Tactics:
- Read the 2 and 3 star reviews on the top 10 competing products. Five star reviews are vague and one star reviews are often shipping rants. The middle is where the useful detail lives.
- Search Reddit and niche forums for "[product] recommendations," "[product] problems" and "anything better than [brand]."
- Tally complaints into clusters. If 40 reviews mention the lid leaking and 3 mention the colour, you know what to fix.
- Write one sentence: "Buyers of X keep complaining about Y, and my version fixes it by Z." If you can't write that sentence, your idea isn't differentiated yet.
Pass if: you've found a complaint that shows up repeatedly across multiple competitors, and your idea directly fixes it.
Rung 3: The smoke test
This is the closest thing sellers have to Bezos's "shorten the distance" goal. You put a product in front of real strangers before it exists, and you measure whether they want it.
How it works:
- Build a simple landing page with a realistic render or mockup, a clear headline built around the complaint you found, a price, and a call to action.
- Run a small paid ad budget (Meta, TikTok or Google) to a targeted audience.
- Measure clicks, time on page and signups.
Keep it honest. The call to action should be "Join the waitlist" or "Get notified at launch," or a button that leads to a clear "We're not quite ready yet, leave your email and get early access" page. Don't take payment for a product you can't deliver, and don't present renders as photos of stock you have on hand.
Tactics that make smoke tests more useful:
- Test two or three angles, not one. Same product, different headlines. "Finally a lunch box that doesn't leak" versus "The lunch box that keeps food hot for 8 hours." The winner tells you what to build and how to sell it.
- Show the price. A test without a price measures curiosity, not demand.
- Use cold audiences only. Your friends, family and existing followers will make any idea look great.
- Spend enough to see a pattern. A few hundred clicks is a reasonable minimum before you read too much into the numbers.
Pass if: you hit the signup threshold you set before the test (see the scorecard below).
Rung 4: The short-form content test
Short-form video is a free focus group with millions of people in it. If your idea can't get attention as content, it'll struggle to get attention as an ad.
Tactics:
- Post 5 to 10 short videos showing the problem, a mockup or sample, and the "what if" moment. Use different hooks for each.
- Watch saves, shares and comments, not just views. "Where can I buy this?" and "Take my money" comments are the signal you want.
- Track which hook drives the most engagement. That's your future ad and your future listing headline.
- If you don't have a sample yet, problem-led content still works. "Why does every [product] do [annoying thing]?" tests whether people feel the pain.
Pass if: at least one or two videos clearly outperform your baseline and you're getting purchase-intent comments from strangers.
Rung 5: Pre-orders and paid waitlists
Signups are cheap. Money is truth. This is the rung where you find out if people will pay your actual price.
Options:
- Pre-orders on your own store. Shopify and similar platforms have pre-order apps. Be crystal clear on the expected ship date, and make refunds easy.
- Refundable deposits. Ask for a small deposit (say 10 to 20 percent) to reserve one from the first batch. Lower friction, still a real commitment.
- Crowdfunding. Kickstarter and Indiegogo work well for novel products with a story. They're a demand test and a funding round in one, but campaigns take real work to run well.
- Email your waitlist first. Offer early signups a founder price for a short window. The conversion rate from waitlist to paid order is one of the most honest numbers you'll get.
Watch the rules. Consumer protection rules on shipping times and refunds apply to pre-orders. In the US, the FTC's mail order rule covers stated shipping times. In Australia, the ACCC and consumer guarantees apply. Check the rules in every market you sell into, and if in doubt, get advice.
Pass if: a meaningful slice of your waitlist converts to paid orders at your target price.
Rung 6: Samples and micro-batches
Now you're spending real money, but still far less than a full production run.
Tactics:
- Order samples from 3 to 5 suppliers. Compare quality, communication speed and how they handle your changes. The cheapest sample is rarely the best partner.
- Negotiate a test MOQ. Many suppliers will do a smaller first run at a higher unit price if you're upfront that it's a trial with a bigger order to follow. Paying more per unit on 100 units beats getting stuck with 1,000.
- Use print-on-demand or local makers where you can. For apparel, printed goods and some accessories, you can test designs with zero inventory.
- Put product in real hands. Send samples to 10 to 20 people who match your buyer, not your mates. Ask them to use it for a week and tell you what's wrong.
- Check compliance early. Safety standards, labelling rules and certifications vary by category and country. Finding out after you've ordered stock is expensive.
Pass if: the product consistently meets spec, testers confirm it fixes the complaint you targeted, and your landed cost still leaves a healthy margin.
Rung 7: The marketplace micro-launch
This is the final exam. Cold traffic, real fees, real competitors.
Tactics:
- Pick one main channel to start. eBay and Etsy are low-cost places to test with small quantities. Amazon gives you more traffic but more cost and complexity. TikTok Shop suits products that demo well on video.
- Ship the micro-batch yourself or through a 3PL before committing to bigger fulfilment programs.
- Track true unit economics. Landed cost, platform fees, ad spend per sale, returns. A product that sells but loses money on every unit is not a winner.
- Get early reviews the legitimate way. Use platform-approved review programs where available and follow-up emails. Never pay for or incentivise reviews outside platform rules.
- Read every return reason and every question. They tell you what to fix before you scale.
Pass if: you're selling steadily to cold traffic at a positive contribution margin, with reviews that confirm the product delivers.
Rung 8: The scale order
You've earned it. By now you know what to make, who buys it, what they'll pay, how to describe it and where it sells. Your first big production order is no longer a bet. It's a decision backed by data.
Even here, don't go all in. A common approach is to order enough for two to three months of your tested sales rate, with a reorder plan ready, rather than a year's worth up front.
The kill, iterate or scale scorecard
The most important part of any test happens before you run it: deciding what result means stop, tweak or go. If you set the bar after you see the numbers, you'll talk yourself into anything.
Here's a starting framework. Adjust the numbers to your category and price point.
| Test | Metric | Kill | Iterate | Scale up a rung |
|---|---|---|---|---|
| Smoke test | Visitor to signup rate | Under 3% | 3% to 8% | Over 8% |
| Smoke test | Ad click-through rate | Under 0.5% | 0.5% to 1.5% | Over 1.5% |
| Content test | Purchase-intent comments | None across all posts | A handful on one post | Repeated across several posts |
| Pre-orders | Waitlist to paid conversion | Under 2% | 2% to 5% | Over 5% |
| Micro-batch | Tester verdict | Core complaint not fixed | Fixed, but new issues found | Fixed, and testers would buy again |
| Micro-launch | Contribution margin after ads | Negative | 0% to 15% | Over 15% |
| Micro-launch | Return rate | Well above category norm | Around category norm | Below category norm |
These thresholds are ShelfTrend rules of thumb to get you started, not industry benchmarks. Higher-priced products usually convert lower, so adjust down for premium items and up for impulse buys.
"Iterate" doesn't mean "keep going anyway." It means change one thing (headline, price, audience, feature) and re-run the same test. If two rounds of iterating don't get you to a pass, kill it and move to the next idea.
Your toolkit by stage
You don't need an expensive stack. Most of this can be done with free or cheap tools you may already use.
| Stage | Tool types | Examples |
|---|---|---|
| Demand research | Search trends, marketplace data, sold listings | Google Trends, eBay sold filter, Amazon BSR tools, TikTok Creative Center |
| Complaint mining | Reviews, forums, AI summarising | Amazon reviews, Reddit, niche forums, an AI assistant to cluster complaints |
| Mockups and renders | AI image tools, 3D mockups, design apps | AI image generators, Canva, Blender, supplier stock renders |
| Smoke tests | Landing pages, ads, analytics | Shopify, Carrd, Squarespace, Meta Ads, TikTok Ads, GA4 |
| Pre-orders | Pre-order apps, crowdfunding | Shopify pre-order apps, Kickstarter, Indiegogo |
| Samples and micro-batches | Low-MOQ sourcing, print-on-demand | Alibaba, 1688 via agents, local makers, print-on-demand services |
| Micro-launch | Marketplaces, small-batch fulfilment | eBay, Etsy, Amazon FBM, TikTok Shop, a local 3PL |
Notice the theme. Tools like AI image generators and AI assistants are doing for product testing what Marketplace and FBA did for selling: collapsing weeks of design and research work into hours, so you can test more ideas for the same money.
A 30-day test plan you can copy
Week 1: Research and pick a lane
- Shortlist 3 to 5 ideas.
- Run desk research and complaint mining on each.
- Write your "Buyers of X keep complaining about Y, my version fixes it with Z" sentence for each survivor.
- Drop anything you can't write that sentence for.
Week 2: Smoke test
- Create mockups for your top 2 ideas.
- Build simple landing pages with 2 headline angles each.
- Write down your kill, iterate and scale numbers.
- Run $100 to $200 of ads per idea to cold audiences.
Week 3: Content and refinement
- Post 5 to 10 short videos around the winning idea and angle.
- Request samples from 3 to 5 suppliers.
- Iterate on anything that landed in the "iterate" band.
Week 4: Ask for money
- Email your waitlist with a founder-price pre-order or refundable deposit offer.
- Compare sample quality and supplier responsiveness.
- Make the call: kill, iterate, or order a micro-batch.
Total spend for most sellers: a few hundred dollars plus samples. Compare that to the cost of 1,000 units of the wrong product sitting in a warehouse.
Red flags that fool sellers
- Friends and family enthusiasm. They love you, not your product. Only cold-audience data counts.
- Vanity metrics. Views and likes feel great but don't pay the bills. Track signups, deposits and sales.
- Testing without a price. Interest at an unknown price tells you almost nothing.
- Moving the goalposts. If you change your pass mark after seeing results, you've stopped testing and started hoping.
- Copying a viral product late. By the time a product is everywhere on TikTok, the margin is usually gone. Test early or test something differentiated.
- Skipping compliance. Electrical, kids, cosmetics, supplements and food-contact products often need specific testing and labelling. Budget for it before you scale.
- Ignoring tariff and freight risk. Duties and shipping costs can swing fast. Build a buffer into your landed cost, and talk to a customs broker for anything complex.
- Overselling pre-orders. Promising a ship date you can't hit damages trust and can breach consumer law. Pad your timelines.
The bottom line
Bezos built Amazon by making it cheaper and faster to try things, and by giving sellers infrastructure they used to have to build themselves. You can run your product pipeline the same way. The tools to test a product idea for a few hundred dollars are already sitting in your browser.
The sellers who win aren't the ones with the best ideas. They're the ones who test the most ideas cheaply, kill the losers fast and only bet big on what buyers have already proven they want.
Your idea probably isn't dying because you lack imagination. It's dying because testing feels too hard. Climb the ladder one rung at a time and it isn't.
Sources:
- Bezos quote: Fox interview, as supplied