Helium 10 Product Research: Bradley Sutton’s Black Box Tips

Why start Amazon product research with keywords?

Bradley’s starting point is simple: look for the language of demand before choosing a product. A keyword can tell you what a customer wants, where they intend to use it, and which feature matters to them.

Compare “kitchen” with “coffee cup holder wall mounted,” an example from the interview. The first is a broad category. The second describes an object, its purpose, and how it should be installed. You can picture a possible product before opening a supplier catalog.

That does not mean every longer phrase converts well or every one-word phrase is weak. It means specific searches can make the customer’s intention easier to understand. Your next job is to see whether Amazon’s existing products satisfy that intention.

Bradley illustrates this with Project X’s coffin shelf. He discusses roughly $200,000 in historical sales and shows the project account during the interview. Treat that as context for the case study, not as a profit figure or a forecast for a new seller.

The useful lesson is curiosity. You do not need to personally want a coffin-shaped shelf to investigate demand for one. Equally, finding an unusual idea does not establish that you can source, differentiate, and sell it profitably.

Set up Black Box Keywords without copying blindly

Open Black Box, select your intended Amazon marketplace, and choose the Keywords tab. The interview uses keywords to discover products, rather than starting with the Products tab. Menu labels may change, so follow the matching controls in your account.

Here are Bradley’s original settings. He repeatedly says these are starting points, not magic numbers. Use them to understand how the filters work, then adjust them to your budget, category knowledge, and tolerance for competition.

The price range is especially easy to misread. An average selling price between $15 and $50 does not mean every result falls within that range. It also says nothing about your landed cost, packaging expense, or minimum order quantity.

The exclusions reflect the timing of the recording. Bradley wanted fewer Christmas-related ideas in that particular search. If seasonal products are your specialty, excluding them automatically could remove opportunities you actually want.

Do not make the search so restrictive that only a handful of familiar products can qualify. If you get very few useful results, loosen one condition and inspect what changes. If the results are mostly irrelevant, tighten the condition connected to that problem. Adjusting the search deliberately teaches you more than replacing every number at once.

Keep a record of each search. Change one or two filters at a time so you understand why a new set of results appeared. A screen full of results is useful only when you can explain what qualified them.

Read competitor revenue and review filters correctly

The most useful part of this demonstration is the move beyond a market average. Bradley asks how many of the leading products meet a particular revenue threshold and how many have relatively few reviews.

His revenue example requires at least five of the top ten products to exceed $2,500 in estimated monthly revenue. This is different from asking whether the average product earns $2,500. An average can hide a single large seller surrounded by weak performers.

His review example asks for at least six of the top ten to have fewer than 150 reviews. That can make a niche worth investigating, because the visible results are not exclusively dominated by listings with enormous review histories.

Helium 10’s official advanced rank-filter guide still documents this count-based approach. It distinguishes competitor counts from averages and explains that the count values operate within a one-to-ten range.

Imagine ten results where five meet your revenue target and six meet your review target. Those two groups could share only one product. In that situation, you have much less evidence of newer listings succeeding than the filter summary might suggest. Your shortlist should identify that overlap explicitly, rather than assuming six lower-review products all have strong sales.

A low review count might describe a newer competitor, an overlooked niche, or a product that barely sells. The filter cannot decide which explanation is correct. Treat the result as a question to investigate, not permission to place an order.

What the unusual product examples actually teach

During the demonstration, Bradley encounters bourbon candles, chocolate sphere molds, beer tasting glasses, ceiling curtain tracks, and wall-mounted coffee cup holders. Their value here is educational: they show how unfamiliar customer searches can lead to concrete product ideas.

With bourbon candles, curiosity comes first. What does the phrase mean, and what do buyers expect? Before comparing sales figures, open the results and understand the actual product. Guessing from a keyword alone can send your research in the wrong direction.

The chocolate sphere mold example prompts a different question: could this already be a crowded or fast-moving trend? Several lower-review listings do not automatically mean you are early. You still need to examine current demand and the strength of the products already selling.

The wall-mounted coffee cup holder is easier to visualize. It also creates practical questions: which wall types can it fit, how many mugs does it hold, and do customers complain about installation? Those questions can become differentiation research.

None of these examples is a current recommendation to source that item. Re-run the search today, inspect the present competition, and validate the customer problem. The repeatable skill is learning how to investigate unfamiliar demand.

Use Xray to see who is actually selling

Once a keyword looks interesting, search it on Amazon in the same marketplace and open Xray through the Helium 10 Chrome extension. Black Box helps generate the lead; Xray helps you examine the products behind it.

The official Xray overview describes on-page analysis of sales, revenue, and pricing. In this workflow, those figures help you compare relevant listings and notice whether demand appears spread across several products or concentrated in one.

In the interview, Bradley sees a result set with one stronger seller and several listings that look much newer. His reaction is optimistic. My practical takeaway is to investigate the weaker listings before assuming their sales are available for you to capture.

A dominant product may have a stronger brand, better fit for the query, better images, or more convincing reviews. A new listing may simply have had little time to establish itself. Neither explanation is obvious from one revenue column.

Also separate relevant products from unrelated results and notice sponsored placements. A product shown on the page is not automatically an organic competitor for the exact need you are evaluating.

Look at brand concentration as well as the number of listings. Several results may belong to the same brand or represent closely related variations. Counting each as evidence of an independent seller succeeding can exaggerate how open the market is. Compare like-for-like products, and note bundles or different pack sizes before interpreting their prices and revenue.

Save a dated shortlist with the keyword, marketplace, relevant listings, and your observations. That makes the next research session a comparison, rather than another round of browsing from memory.

Check title density and relevance together

When I ask Bradley which metrics help validate a product, he highlights a simple manual check: how many page-one listings contain the exact searched phrase in their title?

He contrasts a competitive phrase such as “collagen peptides” with the original coffin shelf discovery. In his account of that early research, only one listing on the first page included “coffin shelf” in its title, and only one was actually that type of shelf.

That second detail matters. The opportunity was not merely missing words. The search results also appeared to offer few products that directly matched what the shopper wanted. Relevance gives the title check its meaning.

Helium 10’s metric definitions describe title density as the number of page-one products with the searched keyword in their title. You can perform Bradley’s basic check manually without making a particular analytics feature the starting requirement.

For your own shortlist, record the exact phrase and count matching titles. Then open the products. A relevant item might use a synonym, a different word order, or a brand-led title while still meeting the customer’s need.

Low title density can suggest a gap worth investigating. It cannot guarantee an easy ranking, a low advertising cost, or a successful launch. Adding the phrase to your title is useful only when the product actually matches it.

Look for listing weaknesses you can genuinely improve

Bradley also notices coffee cup holder images that do not use white backgrounds. In the interview, he treats this as a possible sign that some competitors are less experienced. It is a useful prompt for closer inspection, but an image alone cannot tell you a seller’s overall ability.

Amazon’s main product image requirements generally call for a pure white background, with limited product-type exceptions. Check the guidance for your category rather than treating every non-white image as the same violation.

For research, look beyond compliance. Can shoppers understand the dimensions, installation method, included parts, and intended use? Read customer questions and critical reviews to identify missing information or recurring problems.

Then ask whether you can deliver a meaningful improvement. Clearer instructions, sturdier construction, more suitable dimensions, or better packaging may matter more than a cosmetic change. Write down the benefit in the customer’s language and verify that your supplier can provide it.

Validate the product before ordering inventory

The interview is strongest as a discovery lesson. Before committing to a product, I would add a separate validation pass. These checks extend the method; they are not claims that Bradley proved every example profitable during the recording.

First, review demand over time. A promising current estimate may reflect a holiday, a temporary trend, or a short-lived event. Compare available historical data and related search phrases so one unusually strong period does not become your entire forecast.

Second, investigate the economics using real supplier information. Include product cost, packaging, freight, applicable duties, Amazon fees, advertising, returns, and other relevant expenses. Keep estimates separate from confirmed quotes.

For a simplified hypothetical example, a $30 selling price minus $24 in combined per-unit costs leaves $6 before any omitted expenses and taxes. If advertising or shipping adds another $3, that contribution falls to $3. Revenue filters do not perform this calculation for your business.

Third, test the product itself. Order samples, compare them with the customer needs you identified, and check whether your proposed improvement works. A supplier description is not a substitute for handling the item.

Ask suppliers about the details that could change your decision: sample quality, production lead time, packaging dimensions, minimum order quantity, and the cost of your proposed improvement. An affordable sample does not establish the economics of a full shipment. Put the quote date beside each figure so an old estimate does not silently become a purchasing assumption.

Finally, establish whether you can responsibly sell it in the intended marketplace. Research applicable category restrictions, product requirements, and potential intellectual-property issues before committing. An attractive keyword does not resolve those questions for you.

Turn the method into a repeatable research session

Start each session with one clear constraint, such as a category you understand or a budget you can support. Run a Black Box Keywords search and save a small set of relevant candidates. You do not need dozens of tabs to make progress.

Next, compare the same things for every candidate: customer intent, current competitors, estimated sales distribution, review counts, title relevance, and a possible improvement. Consistent notes make different niches easier to compare.

A simple research note can contain the exact keyword, research date, marketplace, competing product identifiers, the customer problem, and the biggest unresolved question. Add a reason for keeping or rejecting the candidate. This record makes it easier to return later without repeating the entire investigation or forgetting why an initially exciting idea failed your checks.

Finish with a next action. One idea may need a supplier quote; another may need more seasonal data; a third may deserve rejection. Finding a reason to stop is useful research too. The goal is a better decision, not a bigger collection of interesting keywords.

Helium 10 product research: common questions

Start with one keyword and follow the evidence

What I like about this conversation is the combination of curiosity and specific checks. Bradley does not need to recognize every product before investigating it. He starts with customer searches, then asks what the existing listings reveal.

Try that with one keyword. Understand the need, inspect the competition, and record what you still need to verify. If the idea survives those checks, take the next step. If it does not, you have learned something before paying for inventory.

Vova :)