How We Test And Review Software?
The number of features on a software company’s website tells you little about how well its product will work for you. A platform built for creator discovery and one built for affiliate attribution may both be excellent without offering the same tools. We test the software itself, evaluate the work it is designed to do, and explain where its strengths matter to a buyer.
Our process has three connected parts: a feature-based IMH Score on profiles evaluated under this method, a focused assessment of the relevant task in our software comparisons, and video walkthroughs that let readers see the product in use. We have been reviewing marketing tools since 2016, and we use this process to make the reasoning behind our assessments visible.
1. We explore the product before selecting features
Our testing team gets access to the software, often through a demo account provided by the company being reviewed. We test the available features we can access, following the steps a user would take to make them work. We examine the interface, settings, outputs, limitations, and connections between features. A feature appearing in a menu is not enough to establish what it can do.
We map the platform’s intended workflow before assigning ratings. For an influencer marketing product, that could span creator search, audience vetting, outreach, campaign management, content approvals, payments, affiliate tracking, and reporting. A specialist may cover only part of that journey; we assess it against the work it promises to help its intended users complete.
Access can vary. Some functions require a paid tier, connected store or social account, live campaign data, enterprise contract, or managed service. We distinguish a workflow we completed ourselves from one shown in a demonstration, described in current product documentation, or claimed by the company. We do not present an untested capability as a result of hands-on testing.
2. We select ten defining features
Some platforms offer thirty or more features. Scoring every item equally would reward a long feature list and obscure what makes the product useful. After exploring the platform, we select ten features that best explain its core value and strongest areas. Selection is based on five questions:
| Selection question | What it means in practice |
|---|---|
| Does it matter to the intended user? | The feature helps solve a substantial problem for the type of team the product serves. |
| Is it central to the product? | It represents a core workflow or an essential step in delivering the company’s promise, rather than a minor extra. |
| Does it reveal a meaningful strength or difference? | Its execution shows something useful about the platform beyond a standard feature label. |
| Does it matter in day-to-day use? | We consider the work needed to use it, including collaboration, controls, integrations, and scale where relevant. |
| Can its current availability be established? | We check the plan, access requirements, and whether it is generally available, in beta, custom-built, or delivered through a managed service. |
These are the ten most informative features to assess, not a list chosen after scoring to contain only the highest marks. If a core capability has a weakness, it remains part of the assessment. We can discuss additional features in the written review without including each one in the numerical score. A function’s absence or a limitation that materially affects the product’s promise is also explained, even if it is not one of the ten rated items.
3. We test the complete workflow
For each selected feature, we follow a realistic task from input to result. When testing creator discovery, for example, we do more than check whether search filters exist. We enter a brief or search criteria, refine the results, examine creator and audience information, build a shortlist, and see how that shortlist moves into the next step. We assess the relevance of the output, the quality and clarity of the underlying data, and the effort needed to repeat the process.
We also check how a feature behaves beyond the simplest demonstration. Depending on the product, that can involve repeated searches, larger result sets, multiple records or collaborators, bulk actions, complex reports, or exports. We record practical constraints such as usage limits, extra fees, missing permissions, slowdowns, failed actions, or manual work needed outside the platform. The scenario must fit the product; these checks are not a single standardized load test applied to every category.
For data and measurement features, we investigate what a number represents. We look at available explanations of data sources, update frequency, estimates, authentication, attribution, and reporting methods. If the company claims a database size, AI result, fraud reduction, or revenue impact that we cannot verify independently, we identify it as a company claim. We do not turn an unexplained metric into a tested finding.
4. We rate each of the ten features
Each feature receives a score out of five using the same five weighted dimensions. What good performance looks like depends on the feature and its intended buyer. For example, data quality is central to creator vetting; in a content approval workflow, the equivalent concern is whether versions, statuses, and decisions are accurate and traceable. A tool designed for a small team is assessed for its intended operating scale, rather than penalized for not having enterprise functions it never promises.
| Dimension | Weight | What our reviewers examine |
|---|---|---|
| Capability depth | 25% | Whether the feature completes the important parts of its promised task, with useful controls and flexibility. |
| Workflow quality | 20% | The steps from input to usable result, including clarity, repeatability, automation, and avoidable manual work. |
| Data and result quality | 20% | The relevance, accuracy, freshness, transparency, or traceability of what the feature produces, as applicable to its task. |
| Operational fit | 20% | Whether it works at the intended buyer’s scale, including relevant permissions, collaboration, integrations, reliability, and limits. |
| Differentiation | 15% | Whether it delivers a meaningful advantage over credible products solving the same problem, rather than merely including a common function. |
We compare features with products serving a similar buyer and task. A creator marketplace is not marked down for lacking the data architecture of an enterprise intelligence suite, and a routine search filter does not earn an exceptional score just because it exists. We consider what the user can accomplish, which plan or service is required, and whether an apparent advantage survives hands-on use.
We score the five dimensions on a five-point scale and calculate the feature rating as:
We round the result to one decimal place for display. These ranges guide our judgment:
| Score | Interpretation |
|---|---|
| 4.5–5.0 | Exceptional execution for its intended task and buyer. |
| 3.8–4.4 | Strong execution, with a clear advantage in at least some respects. |
| 3.0–3.7 | Capable and competitive, with identifiable trade-offs. |
| 2.0–2.9 | Meaningful gaps, restrictions, or manual steps limit its usefulness. |
| 1.0–1.9 | Minimal or weak execution of an important task. |
A score is an editorial assessment supported by the test, not a claim of laboratory precision. Restricted access or missing information affects what we can conclude. We say which parts were observed directly and which depend on documentation or a company demonstration; an important claim that cannot be substantiated does not receive the same credit as a tested capability.
5. We calculate the IMH Score
The IMH Score is the arithmetic average of the ten feature ratings, with each feature contributing equally. We round the result to one decimal place. We do not add points for the total number of functions listed by the company.
For example, the ten feature ratings displayed on our Modash profile total 43.8. Dividing by ten gives 4.38, which appears as an IMH Score of 4.4 out of 5. The named features and their individual ratings remain visible on the profile so readers can see what produced that number.
Because the ten features are selected around each product’s own proposition, the score measures how well it delivers its defining capabilities. It is not a universal score for every possible marketing task. The written review explains the buyer fit and limitations that an average cannot capture.
Some older profiles still show a previous rating breakdown. Until those products are reassessed using ten feature ratings, the breakdown displayed on their profiles remains the explanation for their current score. We do not describe a legacy score as though it was calculated with this method.
6. We assess the task in each software/platform comparison
A software comparison answers a narrower question than a full profile. If the article is about influencer discovery tools, we concentrate on each platform’s ability to find, evaluate, and shortlist relevant creators. Search precision, audience information, result quality, and the route from shortlist to action matter more in that article than unrelated campaign or payment features.
We compare the products as solutions to the same task, while explaining where their approaches differ. One platform may help a team cover a large creator pool; another may be more useful when audience credibility is the bottleneck. We test and write about the capabilities that bear on the article’s question, and discuss adjacent features when they materially change the workflow.
The score in a comparison’s summary table is the platform’s profile-level IMH Score. It is not a second numerical score calculated solely for that article. The comparison’s written assessment tells readers how the platform performs for the specific use case. A 4.4 overall rating should therefore not be read as an automatic 4.4 for influencer discovery, affiliate management, or any other individual task.
7. We show our work on video
Our testing team also records walkthroughs of individual platforms. Viewers can see the interface, the steps required to use a feature, and the output the software provides. We add these videos to software profiles and, where relevant, to the corresponding reviews in software comparisons. They allow readers to judge whether a workflow looks suitable for their own team, alongside our written analysis.
Videos show the version and access available when they were recorded. A later product update or a different subscription tier may change what a viewer sees in their own account.
Independence, limitations, and updates
Companies we review may provide demo access so we can test their software. Providing access does not determine a rating or guarantee inclusion in a comparison. Our advertising and affiliate policy separates editorial judgments from affiliate links. Reviewers disclose personal or financial interests to the editorial team, which can reassign a review where a conflict exists.
We revisit reviews when changes to a product, its integrations, pricing, or availability affect the assessment. Where testing is limited by account access, missing data, a managed-service component, or information the company has not disclosed, we explain the limitation rather than imply we completed a test we could not perform. Readers should use the individual feature ratings, the task-focused comparison, and any available video together to decide whether the software fits their work.