How to Find Products to Sell Online: A Practical Validation Framework

A product is worth testing only when demand, competition, total cost, supplier reliability, delivery, return risk, and channel fit all work together. Trend data alone is not enough.

A rising search graph or a viral video only proves that people are paying attention right now. It says nothing about whether the unit economics survive real shipping and ad costs, whether a supplier can actually fulfill single-unit dropship orders at the lead time they quoted, or whether the return rate for that category will quietly erase the margin.

This guide covers both halves of that problem for sellers working in a dropshipping model: where to find a credible short list of candidates, and how to score, price, and test each one before a single dollar goes toward inventory you don't have to hold.

Each stage below links out to a deeper Doba guide—on trend and seasonal timing, margin calculation, supplier vetting, and listing automation—so this piece works as a starting reference you can come back to, not a one-time read.

The 8-step process at a glance:

1. Define the customer.

2. Find product candidates.

3. Score five validation axes.

4. Calculate contribution margin.

5. Verify suppliers and listing requirements.

6. Prepare the offer.

7. Run a controlled test.

8. Decide: Test, Continue Research, or Abandon.

Start With the Customer, Not the Product

Before evaluating any product idea, define the target customer across six dimensions — who they are, what problem or use case brings them to this purchase, why they're buying now, what they expect to pay, where they already shop, and how fast they expect delivery. A product idea with no answer for these has nothing concrete to validate against.

The order most retailers get backwards is picking a product first and working out who might buy it second. Flip it: before researching a single product, fill in a short profile.

Fill in this profile before you go looking for a product, not after. Missing an answer for even one of these six dimensions is something you'll end up guessing at later — usually mid-test or mid-launch, when guessing is expensive.

Why Problem-First Beats Trend-First

A trend tells you attention exists right now. A problem, tied to a real customer profile, tells you demand will still exist in six months. Products built around a describable, recurring problem are easier to validate, market, and defend once competitors show up — because your positioning is the problem you solve for a specific buyer, not the object you happen to be selling.

Consider two versions of the same physical object. Version one: "foldable phone stand," listed because phone stands are always popular — no customer profile behind it.

Version two: a foldable phone stand for content creators filming in small apartments without desk space for a permanent tripod, price-sensitive but willing to pay a premium for something that folds flat, likely to buy through a niche Shopify store or TikTok Shop rather than a generic marketplace search.

The object might be identical. But version two comes with all six dimensions filled in, plus a set of complaints — cramped desks, wobbly stands, awkward filming angles — you can go verify are real before you spend a dollar on inventory.

Where to Find Product Ideas

Reliable product ideas come from eight kinds of verifiable signal — cross-checked against each other, not trusted on their own.

Marketplace Bestseller and Trend Pages. Amazon's Movers & Shakers, Etsy's bestseller lists, and eBay's trending categories show what's actually converting into sales right now, not just what's being talked about. They're free, they update constantly, and they're the fastest way to confirm whether interest you've seen elsewhere is showing up as real purchases — a product with attention but no bestseller presence anywhere is a signal to keep watching, not to order inventory yet.

Google Trends. Check whether interest in a category is rising or already peaked over a multi-year window, not just the last few weeks. A term that spikes and fades tells a very different story than one climbing steadily, and comparing a candidate against a couple of related terms shows whether the whole category is growing or just one keyword variant.

Search Queries. Beyond a directional trend line, keyword research tools quantify how big the opportunity actually is and what kind of intent sits behind it. A term with commercial-intent modifiers — "buy," "best," "review" — pulling a few thousand searches a month usually points to real purchase intent; the same volume on a purely informational phrase often doesn't.

Social Platforms. Subreddits, Facebook groups, TikTok, and niche forums where your target customer already discusses the problem are a faster read on real language and real frustrations than any single search tool. Treat a single viral video with caution — it can push attention up for a week and back down just as fast, so weigh it against the marketplace and search signals above before deciding it reflects lasting demand rather than a one-off moment.

Reviews and Customer Complaints. Read the one-, two-, and three-star reviews on products similar to what you're considering. Complaints about sizing, durability, missing features, or confusing instructions are free market research — they tell you exactly where the next version needs to improve.

Supplier Catalogs. Browsing wholesale and dropship supplier catalogs directly — by category, price tier, or shipping origin — surfaces products you won't find by searching for a specific idea, since suppliers often list variations and adjacent items side by side. It also puts you in front of sourcing details (cost, MOQ, lead time) while you're still validating demand, rather than as a separate step later.

Doba Product Discovery Tools. Curated, data-informed catalogs can shortcut the first pass by organizing products around demand signals instead of a raw supplier list. Browse Doba's High Profit Products catalog, for instance, to pull a shortlist that's already organized around margin and demand signals instead of starting from a blank search.

Seasonal and Event Calendars. Timing matters as much as the idea itself for anything tied to a season or holiday. Check Doba's seasonal product calendar for a quarter-by-quarter breakdown of what tends to sell when, and its guide to maximizing Q4 dropshipping success if the holiday quarter is central to your plan — Q4 alone can account for a large share of a store's annual sales, so a strong idea launched at the wrong moment can still underperform a merely good one launched at the right one.

None of these eight sources is conclusive on its own. A rising trend line with no bestseller presence anywhere might mean you're early to a real opportunity — or it might mean nobody has figured out how to source and fulfill it profitably yet. Cross-check at least two of these sources before treating an idea as a real candidate worth scoring against the framework below.

The Product Validation Framework

Score every candidate against five independent axes — demand & competition, differentiation, unit economics, channel fit, and operational compliance — because a product can pass four of these and still be worth abandoning because of the fifth.

According to CB Insights' widely cited analysis of startup failure post-mortems, a lack of market need is the single most commonly cited reason founders give for shutting down, appearing in roughly four out of ten cases. That's the demand axis failing on its own — and it's also the easiest one to check for free before you spend a dollar.

Tools can speed up that check specifically. Doba Pilot's Market Scout, for instance, pulls demand, competition, pricing, and seasonality signals into one report and connects them straight to sourceable products — a faster first pass, though it's still worth working through the rest of the framework yourself rather than treating the report as a final answer.

Doba's broader approach to data-driven product research runs on the same logic: pairing catalog-wide demand signals with expert curation instead of a single trend graph.

The other three axes fail just as often, just more quietly. A product with real demand can still lose on economics once real costs are counted (next section), or on operational difficulty if it's the kind of item that generates a support ticket every time it ships.

A quick way to stress-test differentiation: try describing your version to a friend in one sentence without using the words "cheaper," "better," or "like [competitor] but." If you can't do it, the differentiation axis probably isn't there yet, and that's worth knowing before a listing gets built around it.

Operational difficulty is the axis most often skipped, because it doesn't show up in a spreadsheet the way cost and price do. A product that's fragile, comes in multiple sizes, needs age or safety certification, or draws frequent "is this authentic?" questions creates ongoing work that a margin calculation alone won't capture.

None of that makes a product unsellable — it just means that cost needs to be priced in from the start, not discovered three months in.

Calculate the True Landed Cost

Landed cost is the product price plus inbound freight, duties, and packaging — not the number on the supplier's quote — and it's the starting point for the unit-economics axis above.

Say a supplier quotes $4.00 per unit. That number feels like your real cost. It usually isn't — not yet. Add inbound freight, any import duties, and the packaging it ships in, and the real figure might land at $5.50 or higher. That gap between the quoted number and the landed number is where a lot of "this should be profitable" math quietly falls apart before a single sale happens.

Get Your Terms Straight First

Four terms get used loosely in dropshipping content, and mixing them up is how a "profitable" product turns out not to be:

• Gross profit — selling price minus landed cost, in dollars. Just the cost of the goods themselves.

• Gross margin — gross profit divided by selling price, as a percentage.

• Contribution margin — selling price minus every variable cost tied to that specific order: landed cost, outbound shipping and packaging, payment and platform fees, ad cost per order, and a returns allowance. This is the number that tells you whether an individual sale is actually profitable once the real cost of acquiring and fulfilling it is counted — and it's the number the unit-economics axis above is scored on.

• Net margin — contribution margin minus your share of fixed operating costs (software subscriptions, tools, salaries) for the period, divided by revenue. This is the true bottom line, and in a healthy business it usually sits a few points below contribution margin rather than being a wildly different number.

This framework is written for a dropshipping cost structure specifically: no bulk purchase, no warehousing, the supplier ships each order individually. A traditional inventory-holding model — buying stock upfront, storing it, shipping in bulk — has a different shape: lower per-unit freight from buying at volume, but real fixed costs (warehouse space, insurance, and the risk of unsold stock) .

Markup vs. margin. Markup is measured against cost; margin is measured against price. A product that costs $8 and sells for $40 has a 400% markup but an 80% gross margin. Suppliers tend to quote markup. Your own math should run on margin — specifically contribution margin, not gross margin, once you're deciding whether to proceed.

The Full Cost Checklist

Contribution margin is only as accurate as the cost list behind it, and not every cost belongs at the per-order level. Nine components cover most of what a dropshipping order actually costs, beyond the number on a supplier's quote — the table below also flags which margin figure each one actually affects, since mixing them up is a common source of an overstated margin:

Running these numbers by hand for every candidate gets old fast. A tool like Doba's profit calculator handles four of the components above — supplier price, shipping, platform and payment fees, and advertising cost — and returns what it labels net profit and net margin.

A common rule of thumb is pricing at roughly 3x landed cost or more, precisely because the nine components above eat into that gross number before it becomes contribution margin you actually keep. 

TrueProfit's current dropshipping benchmarks put a good net profit margin for dropshipping specifically at 15% to 25%, based on its analysis of thousands of active ecommerce stores — a useful reality check against a spreadsheet that only shows gross margin.

Pricing With Room to Breathe

Price too close to landed cost and there's no cushion left once the margin killers above show up. A useful habit: price so the product still turns a profit even if only two or three of those five factors hit you at once.

In practice, that usually means pricing well above the bare 3x-landed-cost floor for anything with meaningful ad spend, stacked fees, or category-level return risk — and closer to the floor only for products with low acquisition cost and low return risk to begin with.

Running these numbers by hand for every candidate gets old fast, especially once discounts, taxes, and subscriptions are added into the same total. If margin is the axis you're optimizing hardest for, Doba's guide to finding high-margin products walks through filtering a catalog for it directly.

Verify the Supplier and Listing Requirements

Before inventory money moves, confirm the supplier can actually deliver at the cost and lead time they quoted — and confirm the product is legally sellable on the channel you plan to list it on.

Supplier Reliability

A short set of questions, asked before you commit, catches most of the expensive surprises:

1. What's the landed cost at your planned order size, and at the next tier up? Not just the headline unit price — the number after freight and duties.

2. Can you send a sample from a current production run, not a showroom piece that may not match what actually ships.

3. What's your typical defect rate, and what happens when a unit arrives damaged?

4. What certifications apply to this exact item, and can the supplier share current documents — not just a general claim.

5. What's the real door-to-door lead time, including freight, not just "ready to ship from our warehouse."

6. Have you independently identified a second viable supplier or an alternate SKU for this product? This is on you to find, not something to ask your current supplier to hand over — a supplier has little reason to point you toward a competitor. Search the same marketplaces and catalogs you used to find this supplier, and confirm at least one backup option before your launch depends on a single source.

7. Can you fulfill and ship single units per order, on an ongoing basis, without raising the per-unit cost or dropping priority below your bulk-buying customers? A standard supplier-vetting checklist built for bulk retail usually misses this question entirely — and it's the one that decides whether you can actually dropship this product without holding stock yourself. Get this in writing, not as a verbal assurance.

Warning signs worth taking seriously: a supplier who won't commit to a lead time in writing, a quote that changes once you ask about volume pricing, reluctance to share current certification documents (versus documents from a different batch or a different SKU entirely), or an unwillingness to name a second factory that could produce the same item. No single one of these is automatically disqualifying, but two or more together are worth pausing on before you place an order.

Listing Requirements Checklist

A supplier can be reliable and a listing can still fail to go live, or get suspended after it does, if these nine items aren't nailed down first:

These same requirements are what decide whether an item clears the channel-fit and operational-compliance axes in the framework above.

Platforms that pre-screen suppliers can shortcut part of this work, though they don't replace asking the questions above for your specific item. Check Doba's Premium Suppliers filter, for example, to narrow your shortlist to suppliers that have already cleared baseline verification — and in some cases met additional performance standards on top of that — useful context when you're comparing options, not a substitute for your own check on the specific item you're sourcing.

For a longer walkthrough of the vetting process itself, Doba's guide to securing reliable, vetted suppliers covers similar ground in more depth, with real supplier examples.

Whatever supplier list you're working from, run every one of these questions before money changes hands. A supplier problem is one of the fastest ways to turn a validated product idea into a bad quarter.

Draft Your Listing Before You Test

A landing page or pre-order page is really just a lightweight listing, so you need a working title, description, and a few photos before you can run the test in the next section — not a finished storefront, just something real enough for a stranger to react to.

Write a title and description that lead with the problem the product solves, using the language you already picked up from review-mining and community research earlier in this guide. Pull a workable first set of photos from your supplier's assets for now; original photography can wait until the idea has cleared the small test.

Browse Doba's AI Listing Optimizer to generate a first-pass title and description quickly, then edit it into your own voice before it goes live. A tool can produce a workable draft fast — the specific language your target customer already uses, gathered earlier in this process, is what actually makes an offer convert.

Once an idea clears testing and you're listing for real, the same kind of automation extends further. Doba's listing and order management tools can keep listings, inventory, and pricing in sync across channels, so you're not re-entering the same product data by hand every time something changes.

Run a Small Test Before You Commit

Test with the smallest, cheapest version of a real transaction you can manage, without holding any inventory yourself, following the same seven steps for every candidate so results are comparable across products.

1.Select one to three products. Narrow your shortlist to candidates that already cleared the five-axis framework above, rather than testing everything at once — testing more than a few simultaneously makes it hard to tell which result belongs to which product.

2. Order samples. Get a physical unit from your chosen supplier for your own inspection — quality, packaging, and how closely it matches the listing photos — before any customer sees it.

3. Complete competition and cost verification. Recheck current competitor listings against your differentiation angle, and rerun your landed-cost and contribution-margin math with current numbers. Both can shift in the time between first scoring an idea and actually being ready to test it.

4. Build a small-scale listing. Use the title, description, and photos from the listing-prep step earlier in this guide — enough to look real to a stranger, not a finished storefront.

5. Test with a limited budget or organic traffic. Drive a modest, defined amount of paid traffic to the listing, or rely on organic reach if you already have an audience — either way, set the budget and the time window before you start, not while you're watching results come in. Real orders placed during the test are fulfilled directly by your supplier through normal dropship fulfillment, so nothing gets bought or held as inventory to cover them.

6. Observe clicks, add-to-cart, conversion, complaints, and returns. Track these against your own store's history and expectations you set before the test started, not a number from someone else's screenshot or a generic industry figure — what counts as a strong result depends on your traffic source, your audience, and your category, and only you have that baseline.

7. Scale once your pre-set conditions are met. Move from test to full launch only when the specific conditions you defined before testing — on cost, on conversion, on return rate — are actually met, not when the results merely look promising.

A modest budget and a couple of weeks are often enough for a directional read, longer if the category has a slow purchase cycle or strong seasonality. If you're testing something seasonal, revisit the timing guidance earlier in this guide before you set your test window — running too close to a peak like Q4 leaves little room to react if demand comes in stronger or weaker than expected.

Reading mixed results. If traffic is cheap but conversion is weak, the price or the offer may be the issue rather than the underlying product. If click-through is strong but comments show confusion, the positioning needs work before you can conclude demand isn't there.

Treat a weak first test as information about the offer — the price, the angle, the audience — rather than an automatic verdict on the product idea itself, and route it back to Continue Research rather than an outright Abandon if only one signal came back soft.

When to Reject a Product Idea

Set your Test, Continue Research, and Abandon thresholds — using the same five axes and auto-disqualify conditions from the framework above — before you run a test. Deciding in advance is the only reliable way to beat the pull of an idea you've already fallen for.

Abandon if you see any of these seven signs, regardless of how strong the rest of the idea looks:

• Profit only holds up at an unrealistic selling price. If contribution margin only clears your floor at a price above what your target customer profile said they'd actually pay, the unit-economics axis is passing on paper, not in practice.

• Authorization or compliance can't be verified. If the brand owner won't confirm reseller authorization, or the supplier can't produce current certification documents for a regulated category, the operational-compliance axis has failed outright — it isn't a detail to chase down after launch.

• Delivery cost comes in higher than expected. A one-time landed-cost surprise is something you reprice around; an outbound shipping cost that's structurally higher than modeled erodes contribution margin on every single order, not just the first one.

• Your primary supplier's inventory is unstable. Frequent stockouts or inconsistent fulfillment from your main supplier, with no qualified second supplier lined up, is exactly the single-point-of-failure risk the supplier-reliability questions earlier in this guide are meant to catch.

• Differentiation is insufficient. If you still can't describe the product in one sentence without "cheaper," "better," or "like [competitor] but," that axis hasn't actually cleared — more research won't fix a product with no real reason to exist yet.

• The return reasons can't be fixed by the listing. If customers are returning for reasons a better photo or description can't solve — sizing that doesn't translate online, fragility in transit, a category with structurally high returns — that's a category-level cost to build into contribution margin, not a content problem to iterate on.

• All of your demand evidence comes from a single viral video. A view count or a like count is attention, not verified demand. The demand-and-competition axis needs at least one more independent signal — a marketplace bestseller position, a sustained trend line, real search volume — before it counts as a Pass.

None of these seven improve with more research. They're structural, which is exactly what makes them worth checking first, before you spend time on anything else.

Continue Research if you have one or more Caution scores and no Fail. A crowded competitive field, a thin-but-positive contribution margin, or an unconfirmed lead time are all gaps you can close with one more data point — a competitor teardown, a renegotiated quote, a written lead-time commitment — before you spend money on a live test.

Move to Test only once every axis clears Pass, or an acceptable Caution you've consciously decided to carry forward rather than one you're hoping will resolve itself.

The return-related sign above is worth grounding in real numbers: the National Retail Federation's 2025 Retail Returns Landscape puts the online return rate at roughly 19.3% of sales — call it one in five online orders. 

Return rates still vary a great deal by category in practice, so check current category-level data for your specific product rather than assuming the blended average applies; if your product sits in a high-return category and you don't have a specific plan to manage it, that's a real, ongoing cost to build into contribution margin, not a rounding error to ignore.

None of this is a call to abandon every idea that isn't a clean Pass on all five axes. A single Caution, with everything else strong, usually means Continue Research rather than Abandon — the point of setting thresholds in advance is to make that call with data, not with either fear or attachment doing the deciding.

Worked Example: Applying the Framework Start to Finish

Everything in this section is a hypothetical, illustrative walkthrough. All costs, supplier responses, and test results are assumed for demonstration purposes — not results from an actual Doba test or a real seller's data. The category below is chosen deliberately to keep the compliance side of the walkthrough simple.

The idea: a compact, wall-mountable drying rack aimed at apartment renters without a dryer or balcony space.

Customer profile: target customer is renters in small apartments without in-unit laundry hookups; the problem is nowhere to dry a full wash load without draping it over furniture; purchase motivation is often a recent move or replacing a bulky rack that doesn't fit; assume an expected price range of $25–$45; assume the preferred channel is a niche Shopify store, since the foldable, wall-mounted angle demos well in a short video and benefits from its own product page rather than competing directly in crowded marketplace search; delivery expectation is one to two weeks, which a niche store can set upfront.

Where the idea came from: assume a search-trend check showed steady, not spiking, interest in "small space drying rack" over the past two years, confirmed by a second signal — steady presence in the relevant marketplace bestseller category, not just search attention. The one- and two-star reviews on several existing drying racks repeatedly raised the same two complaints: too bulky to store, too flimsy for a full load.

This also lines up with a broader pattern Doba has tracked in its home organization niche coverage: post-holiday and spring-cleaning clutter reliably drives a recurring spike in this category — a repeating pattern, not a one-off viral moment, which is exactly the distinction the demand check is designed to catch.

Unit economics (assumed figures): assume the supplier's quoted unit price was $5.80. Once real freight and duties were added, the landed cost came to $6.90 — a 19% jump from the quote, a normal-sized gap and exactly why landed cost, not quoted cost, gets modeled.

At an assumed $34.99 retail price:

• Gross profit = $34.99 − $6.90 = $28.09

• Gross margin = $28.09 ÷ $34.99 ≈ 80%

• Assumed variable costs beyond landed cost: outbound shipping and packaging $5.50, payment and platform fees $1.05, an estimated ad cost per order of $11.00, and a returns allowance of $1.05 (total $18.60)

• Contribution margin = $34.99 − $6.90 − $18.60 = $9.49, or about 27%

• Assume a modest allocated share of fixed monthly software and tool costs at this order volume, roughly $0.50 per order

• Net margin ≈ ($9.49 − $0.50) ÷ $34.99 ≈ 26%

For reference, TrueProfit's benchmarks cited earlier put a good net profit margin for dropshipping at 15%–25%. This hypothetical scenario's assumed 26% net margin sits just above that range — in a real case, a result that far above benchmark would be a reason to double-check the ad-cost and returns assumptions rather than take the number at face value, not a reason.

Supplier check (assumed): two suppliers were sampled. One shipped a sample with a wobbly hinge and failed inspection. The second passed inspection, provided current material documentation on request, and confirmed it could fulfill single units per dropship order at the quoted cost.

Separately, an independent search of the same marketplace turned up one alternate supplier carrying a comparable SKU, so the launch wasn't resting on a single source.

Assume the weight-bearing capacity and included wall-mounting hardware were also checked against the sample — a wall-mounted product carries installation and wall-material-fit risk that a compliance certificate doesn't cover, and a rack that can't hold a wet towel load or ships without usable hardware creates the same kind of return a sizing problem would in apparel.

Decision point one — before any test spending: all five axes clear Pass under these assumptions, and the contribution-margin math clears the pre-set floor. That's enough to justify moving to a small test. It is not a claim that the product is profitable — nothing has been confirmed with a real customer yet.

Running the small test (the seven-step process from earlier in this guide, with assumed results):

1. One product selected — this candidate, already scored above.

2. Sample ordered — assume it arrived matching the listing photos, with no defects found on inspection.

3. Competition and cost re-verified — assume competitor prices were still in the $30–$40 range, and the landed-cost math was re-confirmed at $6.90.

4. Small-scale listing built — assume a simple product page went up using supplier photos and a first-pass description drafted with an AI listing tool, then edited into the seller's own voice.

5. Tested with a limited budget — assume $250 spent over two weeks.

6. Results observed — assume roughly 900 visitors, a 6.5% email opt-in rate, a 3.4% pre-order conversion rate, and no complaints logged during the test window (too early in the process for return data to exist yet).

7. Scale decision — assume these results met the thresholds this hypothetical seller had defined for themselves in advance, before the test started.

Decision point two — after the small test: the assumed results cleared this seller's own pre-set thresholds, which is enough to justify moving from test to a limited initial launch — not enough to call the product "profitable."

A two-week test at low volume doesn't show what happens to ad costs at real scale, what the actual return rate looks like once enough orders have shipped, or whether the supplier's single-unit fulfillment holds up as order volume increases.

If any axis or step above had come back Fail — say, the supplier couldn't confirm single-unit fulfillment, or contribution margin came in at or below zero — the call at decision point one would have been Abandon. 

If a step had come back mixed — an eligible category with an unconfirmed lead time, for instance — the right move would have been Continue Research, not a straight yes or no.

Profitability, in the sense of a proven number, only gets confirmed once real sales accumulate over time. This framework's job is to filter out ideas that clearly wouldn't work and to size a safe first test for the ones that might — not to certify any product, including this one, as a guaranteed winner.

Product Validation Checklist

A one-page record to fill in for any candidate once you've worked through the sections above — and the fastest way to answer "why did we decide this" if you check back on the call in a few months.

Customer profile (fill in first): Target customer · Problem or use case · Purchase motivation · Expected price range · Preferred sales channel · Delivery expectations

Frequently Asked Questions

Q1:How much money do I need to validate a product idea before selling it online? 

A basic validation pass — trend research, a landing page test, and a small sample order — can often be run for a few hundred dollars and a few weeks. That's a small cost next to what a full inventory order costs if the idea turns out to be wrong.

Q2:What's the difference between profit margin and markup? 

Markup is measured against cost; margin is measured against selling price. A product that costs $8 and sells for $40 has a 400% markup but an 80% margin — the same numbers, two very different-sounding percentages. Always calculate margin for your own pricing decisions, since that's the number that determines how much of each sale you actually keep.

Q3:How many suppliers should I get quotes from before choosing one? 

At least two, ideally three. Beyond giving you a sense of typical pricing, having a second viable supplier means one factory's delay, quality issue, or inability to fulfill single units doesn't stall your entire launch.

Q4:Is a low-competition product always a good sign? 

Not necessarily. Sometimes it means you've found a genuine gap; sometimes it means other sellers already tried and the demand wasn't there. Check for substitute products people currently use and look for actual search intent before treating low competition as automatically good.

Q5:What's a reasonable profit margin to expect on a new product? 

There's no single right answer, but a common target is 50%+ gross margin before advertising, aiming to land at 10% to 20%+ net once every cost is counted. Below roughly 50% gross, there's often little room left over to spend on getting a customer in the door once everything else is subtracted.