Ninety-four percent of the B2B SaaS homepages we analyzed explicitly state evidence or social proof. Customer logos, case study numbers, review counts, award badges. That information is almost always there.
Explore the data: browse all 100 companies and filter the results in the Homepage Directness Index →
Thirteen percent explicitly state who the product is for.
We analyzed 100 B2B SaaS homepages across startup, growth, and enterprise companies and scored each one on six dimensions: product category, target customer, primary problem, core product, differentiation, and evidence/social proof. The project is called the Homepage Directness Index.
The question behind it is simple: does the homepage state basic information directly, or does it make the visitor infer it?
The chart above is the main finding. Companies are extremely likely to prove they're credible and much less likely to say what they do, who it's for, or what problem it solves.
What We Measured
The sample is 100 B2B SaaS companies: 42 startup, 42 growth, 16 enterprise. Six additional candidates were excluded because their homepages could not be reliably analyzed due to bot-blocked crawls or unusable homepage text. Those exclusions are documented in the methodology.
Each homepage was evaluated on six dimensions:
| Dimension | What it asks |
|---|---|
| Product category | Does the homepage name the category the product belongs to? |
| Target customer | Does it say who the product is for? |
| Primary problem | Does it identify the problem the product solves? |
| Core product | Does it describe what the product actually is? |
| Differentiation | Does it state what makes the product different? |
| Evidence / social proof | Does it include proof of credibility or results? |
Each dimension received one of three classifications:
- Explicit — the homepage states the information directly, in language a reader could point to
- Implied — a reader can infer it from features, context, or adjacent claims, but the homepage doesn't say it outright
- Missing — not addressed directly or indirectly
The evaluation assumes a reader encountering the company for the first time, with no prior brand knowledge, seeing only the homepage text. No context from other pages or outside sources.
Claude performed the structured classification. Each judgment was paired with supporting homepage text. A human review pass checked the automated classifications. The complete rubric and prompts are published on the Homepage Directness Index methodology page.
The Full Results
Across all 100 homepages, the explicit rates by dimension:
| Dimension | Explicit rate |
|---|---|
| Evidence / social proof | 94% |
| Differentiation | 71% |
| Core product | 57% |
| Product category | 39% |
| Primary problem | 14% |
| Target customer | 13% |
Evidence and differentiation are commonly stated directly. Target customer and primary problem almost never are.
The gap between the top and bottom of that list is 81 percentage points. A homepage is seven times more likely to include a customer logo than to say who the product is for.
One important note on what this measures: the study records whether information is explicitly stated, not whether a homepage is good or effective. A page classified as Implied is not a failing page. It may contain detailed, accurate information. The issue is whether that information resolves into a direct statement a first-time visitor can point to.
What Explicit Looks Like: Cal.com and Resend
Three companies scored 6 out of 6 Explicit: Cal.com, Outreach, and Resend. Every measured dimension was directly stated.
Cal.com
Cal.com's homepage directly describes its target audience, names the product category (booking software), lists specific capabilities, and calls out CRM integrations and routing functionality. Its differentiation language is specific enough that the claim could be checked against the actual product.
A first-time visitor does not need to reconstruct what Cal.com is. The page tells them.
Resend
Resend shows an actual API call on the homepage alongside direct language describing what the product does. The visitor does not need to infer from abstract positioning that this is an email infrastructure product for developers. The page says so, and shows the code.
These are not unusually complex products. The directness is a choice about how to communicate, not a function of what the product does.
What Implied Looks Like: New Relic
New Relic scored 0 out of 6 Explicit. All six dimensions were classified as Implied.
That does not mean the New Relic homepage is sparse or uninformative. It contains extensive feature and capability language. The issue is that none of it resolves into a direct statement for the six dimensions.
For example, the homepage describes tracing agent reasoning through code execution and identifying where logic drifts. That communicates a real capability. But a first-time visitor has to infer what broader differentiation claim it represents, what product category it belongs to, and who specifically it's for.
The Directness Index is not a quality ranking. New Relic is a well-established product. Its homepage reflects a communication style common among recognized enterprise software companies, where much of the work is done by brand familiarity rather than explicit statement.
The classification captures what the text says, not whether the company is successful.
The brand-recognition pattern
Several well-known companies sit at or near the Implied end of the dataset:
A plausible explanation is that recognized brands feel less need to restate basic category and customer information because many visitors already know them. Notion's homepage references its reputation as a knowledge base but does not directly state its product category in the way the rubric requires. Figma uses language about helping teams move quickly in the right direction, which may indicate a benefit but does not directly state the primary problem.
The study does not prove that brand recognition causes companies to omit this information. It's an observation about where in the dataset well-known companies tend to cluster.
Differentiation Drops Sharply by Company Stage
The strongest stage-level finding is on differentiation.
| Stage | Explicit differentiation rate | n |
|---|---|---|
| Startup | 88.1% | 42 |
| Growth | 66.7% | 42 |
| Enterprise | 37.5% | 16 |
That's a 50.6 percentage-point spread between startup and enterprise.
Statistical test
We tested all six dimensions across both relevant segment comparisons, for 12 tests total. Fisher's exact test was used throughout because the enterprise sample is relatively small.
For the startup vs. enterprise comparison on differentiation: p = 0.0002. The result remains significant after correcting for multiple comparisons.
Enterprise vs. growth on differentiation: p = 0.072. Growth companies are not statistically distinguishable from enterprise companies on this measure at the threshold the study used.
The strongest defensible conclusion is narrow: startup homepages in this sample explicitly state differentiation substantially more often than enterprise homepages, and that difference is statistically supported.
Core product follows a similar pattern
Core product explicit rates by stage:
- Startup: 69.0%
- Growth: 54.8%
- Enterprise: 31.2%
Core product is statistically significant on its own but does not survive correction for multiple comparisons. It's a pattern in the data, not a statistically established stage effect.
Product category, primary problem, target customer, and evidence all move in a direction broadly consistent with enterprise companies being less explicit. None reaches corrected significance at this sample size.
Do not read the stage findings as evidence that company stage causes messaging to become less direct. The study measures what appears on these pages at one point in time. The mechanism behind the pattern is not established.
A Practical Application
Two dimensions are stated explicitly on very few homepages: target customer (13%) and primary problem (14%). Those correspond to two questions:
- Who is this for?
- What problem does it solve?
Ask those questions of your own homepage. Can a first-time visitor point to a sentence that answers each one? If the answer requires reading between feature lists, interpreting customer logos, or reconstructing meaning from generalized positioning language, the page would likely be classified as Implied under this rubric.
That may or may not matter to you. But the study suggests it's the norm, not the exception.
A Note on AI Search
This research came partly from thinking about how machines understand company websites. Language models encounter homepage text and construct an understanding of what a company is, what it offers, and who it serves. If a homepage leaves that information implicit, an AI system has to infer it, the same way a first-time human visitor does.
That's an interesting question for future research. But the Directness Index did not test AI citation performance. We cannot claim from this dataset that explicit homepages receive more AI citations, rank better in LLMs, or are summarized more accurately. That would require a different experiment.
Explore the Full Dataset
Homepage Directness Index is an ongoing Pinwheel research project.
The full dataset is interactive and lets readers browse all 100 companies, all six dimensions, each classification, and the supporting homepage text used to reach each judgment. Filters by company and stage are available.
The methodology page publishes the complete rubric and classification prompts.