How to Avoid Wasting Your Testing Budget: Analyzing Competitor Reviews Before Launch
Time to read:
8 min
|
Publication date:
March 24, 2026
Most supplement launches follow the same scenario: we look at competitors' creatives — copy their angles — and test. The logic is clear: if it works for them, it will work for us.
The problem is that by doing this, you are not copying the market, but rather their interpretation of the market. And their interpretation could easily be superficial or outdated. As a result, you are testing assumptions. Your budget is spent on validating hypotheses that could have been weeded out in advance.
There is another way: analyze not what the brands are saying, but what their customers are writing.
When This Is Especially Critical
There are situations where review research is especially critical: entering a new market or a new audience segment, or launching a product with a limited testing budget. The smaller the budget — the more expensive every wrong guess becomes.
Why This Works
Formulas are becoming increasingly similar, ingredients are repetitive, and promises are all the same: energy, support, balance, boost... Competition has long since shifted from the ingredient list — to the formulation of the pain point.
The difference between "hormonal support" and "I want to feel like myself again" — is the difference in conversion rates.
When a person writes a review on Amazon, discusses a product on Reddit, or rants on YouTube, they are not thinking about marketing. They are describing a real-life situation: why they bought it, what they were afraid of, what disappointed them, and what actually worked. This is raw demand data — real gold. The recurring phrases and patterns they use — are your future ad angles.
What Is an "Angle"
An angle — is the specific focus of a pain point or desire around which the entire creative is built: the offer, the funnel, and the communication.
The exact same product can be sold through completely different angles: energy, libido, mental clarity, weight management, or the feeling of "I feel like myself again." Choosing the right angle often has a greater impact on the result — than the quality of the creative itself.
Method: How to Analyze Competitor Reviews
1. Where to Find Data
The main sources — are Amazon reviews, Reddit threads, YouTube comments, reviews on competitors' DTC sites, and comments on TikTok and Instagram.
An important point: look beyond the 5★ reviews. 1–3★ reviews — are a source of barriers and fears that the customer never overcame. This is exactly where the most valuable insights are often hidden.
The reviews are full of fakes. How do you spot them?
It is impossible to filter out fakes completely, but learning to spot them — is very doable. Here is what you should pay attention to:
Too polished. A real review is almost always a bit messy: the person recalls specific details, gets sidetracked, or compares the product to something personal. They make typos! If the text sounds like a promotional brochure — that is a red flag.
No specifics. "Great product, helped a lot" without a single detail — is a classic fake template. Real buyers usually mention when they started taking it, how they felt after a week, or compare it to their previous experience.
Account age and review history. A profile created a month ago with 30 reviews across completely different product categories — is a clear sign of a fake. On Amazon, you can check this by clicking on the author's name.
Review spikes. Dozens of 5★ reviews within one or two weeks — compared to a usual rate of 2–3 reviews a month — is a sign of paid reviews. Especially if everything goes quiet again afterward.
Identical phrasing. If several reviews repeat the exact same phrases almost word-for-word — they are likely written from a template. Especially if it is a direct quote from the product packaging.
Only positives without any nuances. Real buyers almost always mention at least one minor drawback: "the pills are a bit large," "the smell is specific," or "it did not work right away." A perfect review without a single "but" — is suspicious.
In practice, one or two signs do not prove anything yet. But if several overlap — it is better to leave that review out of your analysis.
2. What Volume Is Sufficient
To spot stable patterns, you need hundreds of reviews, not just 50. In practice, 2,000 — 3,000 reviews per category provide a stable picture of frequencies and recurring themes. Any less — and there is a high risk of skewed data.
3. What to Track During Analysis
It is important not just to understand "what they are writing about," but to track specific parameters:
frequency of problem mentions
emotional intensity
context — whether it is an expectation, fear, or disappointment
purchase trigger
purchase barrier
recurring phrasing
Here is an example of what this looks like:
A table like this immediately reveals — where the hidden demand lies.
How to Speed Up the Process with AI
You do not need to read thousands of reviews with your own eyes — you are not masochists. Here is how we automate the process. While other agencies pay for parsers that collect messy data, crash at the worst possible moment, and require a dedicated person to set up — we got tired of that and built our own tools. Our developers built them specifically for our tasks. We use them daily with our clients, and we no longer depend on third-party bugs or pricing policies. But if you want to replicate this yourself, here is the roadmap:
Step 1. Collect the reviews in Google Sheets: one review — one row. Include source, rating, and text. Parsers like Octoparse will speed up the collection, but you can also do it manually, especially if the category is small.
Step 2. Break the data into batches of 50 — 100 reviews and feed them into ChatGPT or Claude with a specific prompt:
Identify recurring formulations of pain points and desires. Group them by theme.
For each theme, rate the emotional intensity: low / medium / high.
Find purchase triggers — what was the final straw before the person made the purchase.
List the barriers — the fears and doubts that people mention.
Create a glossary: actual phrases used by the audience, free of marketing jargon.
Step 3. Consolidate the batch results into a single spreadsheet and look for repetitions. What pops up again and again across different batches — is your stable pattern. That is where you should focus.
4. How to Turn This Into Hypotheses
If a phrase is repeated dozens of times — it is a potential angle. For example, instead of the abstract "hormonal balance support," you see real-life phrases:
"I don't feel attractive anymore"
"my body feels like it's not mine"
"tired of being tired"
"there used to be more desire"
This is the language of the market. An angle is built from the phrases people use to describe their lives — not their illness.
Case Studies from SIRKA
Menopause
The standard set of angles in this category includes hot flashes, weight gain, and mood swings. This is exactly how most brands position their products.
Our analysis of Reddit and YouTube showed something else: the topic of low libido is mentioned less frequently in total numbers, but with very high emotional intensity. Women bring it up themselves, unprompted, with a clear sense that it is a topic people avoid talking about openly.
This angle proved to be the strongest in testing. Not because it was provocative, but because it accurately hit a real, raw pain point that competitors were completely ignoring.
Testosterone Support
The expectation was obvious: libido is the main driver for the male audience in this category. That is exactly where competitors focus their messaging.
Actual review analysis showed otherwise: fatigue and lack of energy are mentioned about 2.5 times more frequently. Shifting the creative priority from libido to energy completely changed the test results.
Sometimes an insight is not a new angle — it is just the correct prioritization of already known ones.
The Bottom Line
Most brands test hypotheses. Those who do research — test pre-filtered hypotheses. The difference lies in how much budget is wasted finding out what doesn't work.
Review analysis is not a magic pill. But it filters out the weakest options before you spend a dime on testing, and occasionally uncovers the gold your competitors completely missed.
