Frequently Asked Questions

Stratify is a B2B contact discovery and profiling platform that discovers professional profiles from public web data using Google search, scores them with a proprietary composite model (CPSEM), and lets you export qualified contacts. It's designed for sales teams, recruiters, and market researchers who need to identify and prioritize high-value contacts.

From the Dashboard, fill in the search parameters on the left panel: role keywords (e.g. "Product Manager"), industry keywords (e.g. "SaaS, Fintech"), and location (e.g. "San Francisco"). Set the number of contacts you want and click the initiate extraction button. Results will stream in real-time as they're discovered and scored.

Each unique profile returned by a search counts as one contact against your daily quota, regardless of whether it came from Google search or from Stratify's cache. Your daily contact limit depends on your subscription plan.

Stratify first checks its internal cache for profiles matching your criteria. Cached results are returned instantly. For new searches, Stratify queries Google's search index using optimized queries built from your role, industry, and location parameters. Each result is then parsed, deduplicated, and scored through the CPSEM model.

A cache hit means the profile was found in Stratify's internal database from a previous search, so no Google API call was needed. The cache hit ratio shows what percentage of your results were served from cache. Higher cache hit ratios mean faster results and better resource efficiency.

The Source column indicates where a profile was retrieved from. 'cache' (shown in green) means it was found in Stratify's database from a previous search. 'google' means it was freshly discovered from Google search results during this job.

Yes. Role keywords, industry keywords, and location all support comma-separated values with OR logic. For example, entering "Product Manager, Engineering Lead" for role will match profiles containing either term. The scoring system evaluates each comma-separated value independently and takes the best match.

The company size band filter lets you narrow results by approximate company headcount. Options include ranges like 1-10, 11-50, 51-200, 201-500, 501-1000, and 1000+. This filter is applied during the search query construction phase to bias results toward companies of that size.

CPSEM stands for Company Profiling & Signal Extraction Module. It's Stratify's proprietary composite scoring engine that evaluates every discovered profile across five dimensions to produce a single 0-100 relevance score. The score helps you quickly identify which contacts are most likely to match your ideal customer or candidate profile.

The composite score is a weighted average of five individual scores: Target Density (20%), Seniority (20%), Industry Alignment (25%), Company Tier (10%), and Market Signal Consistency (25%). Each sub-score ranges from 0 to 1, and the final composite is expressed as a percentage (0-100).

The weights reflect how much each factor contributes to the final score. Industry Alignment and Market Signal Consistency carry the most weight (25% each) because they best predict whether a contact is genuinely relevant. Target Density and Seniority each contribute 20%, capturing role fit and decision-making authority. Company Tier has the lowest weight (10%) since company prestige alone doesn't determine contact quality.

Target Density (weight: 20%) measures how closely a profile's role title and associated text match your search keywords. It answers the question: "Does this person actually do the job I'm looking for?"

It uses token-based overlap matching with two components: Role Overlap (70% of the score) compares the profile's job title against your role keywords. Snippet Overlap (30%) compares the Google search snippet against your industry keywords. The formula is: Target Density = 0.7 × role_overlap + 0.3 × snippet_overlap.

Token overlap is a text matching method. Both your search term and the profile text are split into individual words (tokens), lowercased, and then compared. The overlap ratio is the number of matching tokens divided by the total tokens in your search term. For example, if you search for "Senior Product Manager" and a profile title is "Product Manager at Acme", 2 out of 3 tokens match → overlap = 0.67.

Seniority (weight: 20%) evaluates the professional seniority level of each profile based on their job title. It helps you prioritize decision-makers and senior professionals over junior staff.

Stratify uses pattern matching on job titles to classify profiles into four tiers: C-Level (score: 100), including CEO, CTO, CFO, Founder, Co-founder, Chief officers. VP/Director (score: 75), including VP, Vice President, Director, Head of, SVP, EVP. Manager/Lead (score: 50), including Manager, Lead, Supervisor, Team Lead, Principal. Individual Contributor (score: 25), including Specialist, Analyst, Engineer, Developer, Consultant, Designer, Architect. If no pattern matches, a default score of 30 is assigned.

Industry Alignment (weight: 25%) measures how well the profile's industry matches your target industry keywords. It carries the joint-highest weight because industry fit is one of the strongest indicators of contact relevance.

Stratify uses a two-tier approach. Primary: If an LLM (Gemini) has extracted a structured industry label for the profile, it compares that directly. A match returns a score of 100, a mismatch returns 10. Fallback: If no LLM-extracted industry is available, it uses token overlap between the profile's text and your industry keywords, expanded with industry cluster synonyms (e.g., searching "fintech" also matches related terms like "payments", "banking tech").

Industry clusters are groups of related terms that Stratify uses to expand matching. For example, the "fintech" cluster includes terms related to financial technology, the "tech" cluster covers software and SaaS-related terms, and "healthcare" covers medical and health-tech terms. This expansion ensures that relevant profiles aren't missed due to terminology differences.

Company Tier (weight: 10%) classifies the profile's company into one of three tiers based on company prestige, size, and market presence. It helps you understand whether a contact works at an enterprise, mid-market, or early-stage company.

Tier 1 (score: 100): Fortune 500, FAANG, major global enterprises, and well-known companies across tech, banking, consulting, and industry. Includes companies like Google, Microsoft, JPMorgan, McKinsey, Reliance, TCS, etc. Also matched by patterns like "Fortune 500", "NYSE", "NASDAQ", "S&P 500". Tier 2 (score: 70): Growth-stage and mid-market companies. Matched by patterns like "Series B-F", "mid-market", "scale-up", "unicorn", or headcount indicators (200-500 employees). Tier 3 (score: 40): Startups and early-stage companies. Matched by "startup", "seed", "Series A", "bootstrapped", "small business", or small headcount. Unknown (score: 50): when no pattern matches.

Market Signal Consistency (weight: 25%) is a cross-validation score that checks whether different data points about a profile tell a coherent story. It penalizes profiles where the role, industry, location, or company data contradict each other, and rewards profiles where everything aligns.

The score uses a bonus/penalty system. Bonuses: +30 if the role matches target role keywords, +25 if company/snippet matches target industry, +25 if location matches target location (+15 if only found in snippet), +20 if the company name appears in the search snippet (validates authenticity). Penalties: -20 if role doesn't match and snippet doesn't mention the role, -40 if target location is specified but there's a clear mismatch, -25 if location is specified but the profile has no location data. The final score is clamped between 0 and 100.

Web-scraped data can be noisy. A search for "Product Manager in San Francisco" might return a Software Engineer in New York who happened to appear in search results. Market Signal Consistency catches these mismatches by cross-referencing all available signals, ensuring that high-scoring contacts genuinely match your criteria across all dimensions, not just one.

On the Job Detail page, you can filter by: Min CPSEM Score (only show contacts above a certain composite score), Company Tier (filter to Tier 1, 2, or 3 companies only), and Min Seniority Score (only show contacts above a certain seniority level). Filters are applied in real-time and affect both the table view and CSV export.

There are two ways to export: From the Dashboard, click the "Export CSV" button in the results area to download the currently visible profiles. From the Job Detail page, click "Export CSV" to download the filtered results. The CSV includes all profile fields: name, position, company, location, website, industry, all five CPSEM sub-scores, company tier, source, and profile URL.

The CSV export includes 14 columns: Name, Position, Company, Location, Website, Industry, CPSEM Score (composite), Seniority Score, Density Score, Industry Alignment, Market Signal, Company Tier, Source (cache/google), and Profile URL. All scores are expressed as percentages (0-100).

Stratify discovers profiles exclusively from publicly available web data indexed by Google. It does not scrape private databases, bypass login walls, or access any non-public information. All data comes from public web pages that Google has already indexed.

Fresh search results are fetched from Google's index in real-time. However, Google's index itself may have some lag, so the search results reflect whatever Google has most recently crawled and indexed. Cached profiles in Stratify's database reflect the data from when they were first discovered.

The 'View' link opens the original public web page where the profile was discovered. This could be a LinkedIn public profile, a company team page, a conference speaker listing, or any other publicly indexed page. It lets you verify the data and access additional context.