230: How An 8 Figure Business Launches New Products On Amazon With Bernie Thompson

Podcast: Download (Duration: 53:57 — 123.7MB)

Bernie Thompson plans to lose money on every product launch. Not as a risk he accepts, as a line item in a written business plan: a target ACoS at roughly double his margin, held for weeks or months, treated as the sunk investment required to reach page one.

Bernie runs Plugable, an eight-figure business selling USB and Bluetooth devices, and previously managed the USB and Bluetooth teams at Microsoft. He also built Efficient Era, the Amazon toolset his company developed in-house and now sells, which I have used for my own Amazon business for years.

Below is how launching on Amazon has changed, the business plan he now writes for every product, and why ad spend is really a mechanism for training Amazon’s search engine.

Key takeaways

• Plugable carries 120 to 130 products and launches two to three a month.
• Every product now gets a written business plan modeling keyword run rates and planned losses.
• Launch ACoS target is roughly double the product’s margin, held until the product gains traction.
• Four to eight weeks of data is the minimum before you know whether a launch is working.
• Two months of losses without keyword progress signals failure.
• Ad spend trains Amazon’s search engine about what your product is relevant for, which drives organic position.
• Plugable’s electronics benchmarks: roughly 50 impressions per click and 15 clicks per sale.
• If a keyword shows almost no impressions, that is an indexing problem rather than a bidding problem.

How launching on Amazon has changed

Bernie started in 2009, and describes 2009 to 2012 as a gold rush. A decent product at a decent price with a good listing produced a strong hit rate, with roughly every other product becoming a significant success.

There was no single breaking point. It has simply tightened every year for about five years.

The compounding factors: mastermind groups teaching Amazon as a fast route to starting a business, then Chinese sellers arriving three or four years ago, many of them sitting close to factories.

For electronics specifically there is a fourth factor. Amazon Basics launched around the same time Plugable did, and it works its way to the top of every high-volume category eventually.

Why he still sells only on Amazon

Plugable does not sell on its own site, and Bernie says he sometimes kicks himself over it.

His framing of the choice: send traffic to a marketplace where it spins up a flywheel and generates further sales, or capture it yourself and build an independent customer base.

He continues choosing the flywheel.

The business plan he now writes for every product

Plugable used to run a simple model covering input costs, Amazon fees, and price. That has become a fuller business plan.

The core question is which keywords the product can realistically rank for, including both broad category terms and the specific differentiators of that product.

Then: what run rate is required to reach first page or first position on those keywords, and can the company afford the inventory and the losses required to get there.

The category he walked away from

Bernie’s concrete example is simple USB chargers, a mature category where Plugable would compete directly with Anker.

The top seller in that category runs tens of thousands of units a month. Modelling it out, reaching the top would require roughly a quarter million dollars sunk into inventory alone.

They declined. The probability of success did not justify the fixed investment, and they are waiting for a technology inflection point instead.

His broader point is that every seller has a ceiling on how large a category they can realistically attack, and the business plan is what surfaces that number before you commit.

The risk sitting underneath all of it

Sales velocity in units has been the dominant factor in Amazon search position, measured across overlapping windows of roughly one, three, seven, thirty, and ninety days.

Bernie flags the exposure directly. There has been discussion that Amazon may shift to revenue-based velocity in some categories, and Amazon typically tests changes category by category first.

If that happens, deep discounting to drive unit velocity stops working. You would have built an expensive business model on a platform behavior that changed.

Why he plans to lose money on every launch

Bernie’s launch ACoS target is roughly double his margin. On a product netting 10%, he will run at 20% ACoS and lose money on every sale deliberately.

The reasoning is the same as a giveaway. Each ad-generated sale is not just that sale, it improves search position, which produces organic sales that follow.

There is also weak customer acquisition happening. Amazon controls the customer relationship and prohibits remarketing, and some stickiness still exists and varies by category, so the lifetime value of that customer exceeds the single transaction.

That is why he argues you often want to advertise at an ACoS above your product margin.

Why reviews make early spend more expensive

A new product with no reviews converts worse, so the same ad spend produces fewer sales.

Bernie states plainly that Plugable has never paid for reviews or engaged in review manipulation. The consequence is patience and higher early spend, because generating the first few hundred sales to earn the first handful of reviews is genuinely uphill.

How long to hold the loss

There is no fixed period. Four to eight weeks of data is the minimum before you know anything.

At two months of losses without progress, they treat it as a failure scenario and start scrambling for options beyond ad spend.

The success criterion is not conversion rate. It is whether the product is executing against the business plan: making solid progress toward target position on the identified keywords.

For a failed product, the painful part is the inventory. Plugable ends up with far too much and unloads it slowly, weighing whether to spend on ads to move it or simply cut price.

Which keywords to target first

Bernie’s structure separates two types of keyword, and the sequence matters.

The differentiator keywords are the specific characteristics that made the product worth launching. Someone searching those terms should find your product genuinely compelling.

Broad terms produce poor impression-to-click ratios early, because you are one product in a sea of alternatives and nothing distinguishes you in that result set.

Bernie’s guidance is being honest with yourself about why someone would actually buy your product, then spending energy on those specific characteristics. That is usually where initial traction comes from.

If those differentiators turn out to matter to a large enough market, you eventually climb the broad terms too. If they only appeal to a few people, you have a modest success rather than a category winner.

Tools he uses for keyword research

Efficient Era does not cover keyword research, so Plugable uses Helium 10 primarily, with Jungle Scout, Viral Launch, and Sellics also in the mix.

How Amazon sponsored products actually work

Sponsored product ads are the largest share of Plugable’s launch spend in dollars, though not in time.

Bernie’s argument for them is control. Costs rise as competition increases, and you can control spend precisely and get your product in front of people immediately.

Match types and how keywords get promoted

Amazon offers broad, phrase, and exact match. Moving a keyword toward exact shows your ad to fewer people and requires a lower bid, because an exact match only fires on that precise phrase and tends to win the underlying auction better than a broad match.

The useful mechanic is that Amazon returns search term data. Advertise on “shirt” and the data comes back with every phrase people actually searched containing it: white shirt, shirt for the Broncos, and so on.

That lets you find high-converting specific searches and promote them into their own targeted keywords.

Why he still runs automatic campaigns

Plugable runs both automatic and manual campaigns on every product, and the reason is not primarily keyword harvesting.

Automatic campaigns are the only way to access certain placements. The example Bernie gives is the page shown after checkout completes, where Amazon recommends further products, which he is not aware of any way to reach from a seller account except through an auto campaign. Certain placements on competitors’ product pages work the same way.

What to do when a keyword gets no impressions

This is the failure students hit constantly, and Bernie’s diagnosis is important: if you are not indexing organically for a keyword, bidding on it accomplishes nothing.

You can raise the bid indefinitely and Amazon will return zero impressions or a handful.

The fix is not an advertising fix. Check whether your product actually returns for that term, and if it does not, solve the indexing problem through backend keywords or listing changes first. Once indexed, you can win impressions at lower bids.

Benchmarks and what they actually tell you

Plugable’s electronics numbers: roughly 50 impressions per click, 15 clicks per sale, and about 1,000 impressions per sale. That is a click-through rate near 2% and a conversion rate under 10%.

Bernie is direct that these ratios do not drive his bidding decisions. Everything keys off ACoS, set at the account level and overridden per product.

The ratios matter for a different reason. When they run badly, the corrective action is improving the listing rather than adjusting bids: better images, better text, tighter keywords, and making the product more specifically match what that audience wants.

The hidden cost of bidding too broadly

Overbidding on broad keywords drives the wrong traffic, which costs money and, in some categories, costs you reviews.

If the wrong person buys anyway, that is a prime candidate for a negative review.

Bernie’s observation is that this aligns everyone’s incentives. You want buyers who are excited when they see the product and happy after they receive it, and keyword refinement plus bid adjustment is how you achieve both profitability and relevance simultaneously.

Ads as a mechanism for training Amazon’s algorithm

This is the reframe that matters most in the conversation.

Ad spend is not only about the people who buy from the ad. It is a system for spending money to teach Amazon’s search engine what your product is relevant for, so that organic search starts surfacing you for those terms.

That is why relevance in your keyword selection compounds beyond the immediate return.

The measurement problem neither of us has solved

I raised the obvious question: if losing money on ads spins the flywheel hard enough, you may make more organically than you lose, and breaking even on ads might actually be leaving money on the table.

Bernie agreed and said Plugable has not figured it out and their tools do not yet model it. He knows the ad-generated sale carries both flywheel effects and customer lifetime value, and turning those into numbers would substantially improve the decision.

My own version of this is Facebook ads where the direct numbers do not work, and turning them off causes other channels to decline alongside them. The correlation is clearly real and genuinely hard to measure.

When to shift from launch mode to profit mode

The signal is keyword position rather than a date.

For differentiator keywords Bernie wants position one, not just page one. For broad category terms, page one is often the realistic ceiling.

His reasoning for fighting hard for first position: Amazon is a winner-take-most market with a big head and a long tail, and moving from second to first position can double sales in some categories.

Once the product holds its target positions and ACoS has settled, they drop the target ACoS into profitable territory. Otherwise the tooling keeps bidding up to hit the launch-mode target indefinitely.

What changed in how Plugable markets

Bernie describes the company as historically anti-marketing, on the theory that the technology sells itself and they support products well. That looked cool early and foolish a few years ago.

What stayed consistent is content. A clear brand and message around USB connectivity, extensive compatibility information, and heavy investment in content marketing and video since 2009.

What changed is everything else. They hired a PR agency for the first time in the company’s existence, and built a dedicated marketing team after having zero marketing specialists two years prior.

Their advertising was erratic early, mostly Google AdWords, and they moved that budget to sponsored products within about a year of its launch after seeing far higher effectiveness per dollar.

On press releases specifically, Bernie considers them in decline for 10 to 20 years with little press remaining to pick them up. What does work in tech is one-to-one relationships with well-informed journalists when you genuinely have something new.

What launch promotion looks like now

The core activity is correctly educating the market: being honest about real differentiators that an informed buyer would agree matter, and spending energy highlighting those.

Channels are YouTube, blog posts that double as email blasts to their list, and selective outreach to tech press.

The time cost of managing ads

Before tooling, Plugable had a full-time person spending two to three days a week purely managing ads.

With tooling that dropped to three or four hours. Bernie manages his own account by reviewing hundreds of individual recommendations weekly, accepting or rejecting each, and rejecting mainly on compatibility grounds specific to his products.

Across their whole tool they manage over 10,000 ad campaigns and millions of keywords.

You can find Bernie’s products at Plugable and his Amazon toolset at Efficient Era.

Frequently asked questions

What ACoS should you target when launching on Amazon?

Bernie Thompson targets roughly double his product margin during launch, deliberately losing money on each sale. The reasoning is that each ad sale improves search position, which generates organic sales and carries customer lifetime value beyond the transaction.

How long should you lose money on a new Amazon product?

Four to eight weeks minimum before you have enough data to judge. At two months without progress toward your target keyword positions, treat it as a failure scenario.

How do you know if an Amazon launch is working?

By keyword position against your business plan rather than by conversion rate. You should see progress on your specific differentiator keywords first, and only later on broad category terms.

Why is my Amazon keyword getting no impressions?

Because you are probably not indexing organically for that term, in which case bidding higher achieves nothing. Check whether your product returns for that search, and fix indexing through backend keywords or listing changes before returning to ads.

Should you run automatic or manual Amazon campaigns?

Both. Automatic campaigns are the only way to access certain placements, including the post-checkout recommendation page and some competitor product page placements, beyond their usefulness for keyword harvesting.

What is a good click-through and conversion rate on Amazon ads?

Plugable sees roughly 50 impressions per click and 15 clicks per sale in electronics, so about 2% click-through and under 10% conversion. Bernie notes these vary heavily by category and do not drive bidding decisions, which key off ACoS.

Do Amazon ads affect organic rankings?

Yes, indirectly and substantially. Ad-generated sales drive the velocity that determines search position, and the keywords you advertise on teach Amazon’s search engine what your product is relevant for.

Should you target broad or specific keywords first?

Specific differentiator keywords first, aiming for position one, since broad terms produce poor click-through when nothing distinguishes you in a crowded result set. Broad category terms become realistic only after the differentiators win.