Profound vs Searchable: We Tracked the Same 10 Prompts in Both

Profound and Searchable saw almost the same competitive market in our test, but scored it differently. Across the same 10 prompts, three AI engines, and daily runs, 9 of the top 10 brands overlapped, rank correlation reached 0.88, and top-10 Share of Voice was 66.8% versus 67.5%. Yet Searchable's Visibility Scores ran 3.5 percentage points lower on average.

That gap matters because Visibility Score is the number most teams put into reports. Our data suggests it is useful for tracking movement inside one platform, but not for comparing performance across platforms. The rankings and Share of Voice transferred cleanly; the proprietary score did not.

This comparison goes beyond feature lists. We used the same prompt set in both tools and compared the outputs across visibility, citations, query fanouts, competitive intelligence, crawler analytics, agents, and pricing. The result is less about which dashboard has more features and more about which platform gives you the better operating model for AEO and which metrics you can actually trust when you switch tools.

We Tracked the Same 10 Prompts in Both Tools for Five Days

Both accounts ran the same setup: the same 10 prompts, tracked daily across ChatGPT, Google AI Overviews and Perplexity, at 30 answers per day. Whatever the two dashboards disagree about is the tooling rather than the category.

Below are the two numbers both platforms lead with, Visibility Score and Share of Voice, for every brand reaching either top ten.

The short version: they agree on the landscape and disagree on the scores attached to it.

How we ran it

This is a prompt-set test, not a brand test. Five days at 30 answers a day comes to 150 answers per platform, three responses per prompt, one per model.

Hold on to that number. It matters again at pricing, because it means the two platforms were doing identical work on identical inputs. Any difference in what comes back is the tooling talking, not the market. We changed nothing else.

The full prompt set:

1. What are the best AI video generators?
2. What are the best AI YouTube video generators?
3. What are the best AI TikTok video generators?
4. What are the best AI short video generators?
5. What are the best AI image-to-video generators?
6. What are the best AI tools for video creation?
7. What are the best AI text-to-video platforms for marketing videos?
8. What's the best text-to-video production tool for AI video creation?
9. Which AI video creation company has the best text-to-video production service?
10. Which provider offers the best text-to-video production in AI video creation?

Where the two tools label the same company differently, we harmonised the names for the tables below. That labelling difference turns out to be one of the more interesting results in the whole test, and we come back to it.

Visibility Score

Across the nine brands both platforms scored, Searchable ran 3.5 points lower on average, with an average absolute difference of 5.7 points.

Share of Voice

Share of Voice barely moved. Average absolute difference: 0.67 points, against 5.7 for Visibility Score.

What agreed

• 9 of 10 brands overlapped. Profound surfaced Adobe, Searchable surfaced Canva. Everything else matched.
• Rank correlation of 0.88 across the nine shared brands, on Visibility Score and again on Share of Voice. The two metrics order the brands almost identically inside each tool, so treat this as one result rather than two independent confirmations.
• Near-identical market concentration. The top ten brands account for 66.8% of Share of Voice in Profound and 67.5% in Searchable. Both tools see a category with the same shape and the same long tail.

Two platforms, two proprietary metrics, two independent extraction pipelines, and a competitive picture that is functionally the same. The disagreement sits in the scoring layer on top of it.

Why Their Visibility Scores Disagree by 3.5 Points but Their Rankings Don't

A 3.5-point gap sounds like disagreement about the data. It isn't. Divide every brand's Visibility Score by its Share of Voice and both tools produce a near-constant ratio, just at different constants. That single difference explains most of the gap, and neither vendor publishes it.

For Profound the stability is expected. Its Visibility Score is the share of responses containing your brand and its Share of Voice is your share of all brand mentions. Divide one by the other, the brand-specific terms cancel, and what remains is identical for every brand by construction. Profound's 6.63, with a 2.4% spread, is that identity plus rounding.

Searchable is the anomaly. Its Visibility Score is documented as a composite of mention frequency, position, context quality and sentiment, which has no reason to hold a fixed ratio to Share of Voice. Ours held at 6.05 with a coefficient of variation of 0.57%, four times tighter than Profound's presence-based metric.

Rescale Searchable by 6.63 ÷ 6.05 and the average signed gap moves from -3.5 points to +0.7, with average absolute difference falling from 5.7 to 4.6. Most of the headline disagreement is scaling.

What the documentation says, and what the numbers did

Searchable's docs define Visibility Score as a 0 to 100 prominence measure built from those four factors, and separately define Presence Rate as the percentage of tracked prompts mentioning your brand at all. Weight position and sentiment heavily and the ratio to Share of Voice should scatter. It did not.

Three brands tied at the bottom of Searchable's own leaderboard:

CapCut ranks three and a half positions higher than Canva and carries better sentiment. Under a formula weighting those factors, they should separate. All three land on exactly 17% and 2.8%.

The second signal is sharper. Every prompt-level Visibility Score in the account landed on an exact twelfth: 0%, 17%, 25%, 33%, 33%, 17%, 58%, 17%, 50% and 83%, or 0, 2, 3, 4, 4, 2, 7, 2, 6 and 10 out of 12. Sentiment and average position on that same screen are continuous, running roughly 64.3 to 92.9 and #4.0 to #6.0. Scores that keep landing on exact twelfths fit a discrete count of appearances better than a figure materially moved by those inputs.

We cannot infer a proprietary formula from one dataset. But in our sample Searchable's displayed Visibility Score behaved far more like a presence metric than like a score materially weighted by position and sentiment. Before quoting that number to a client as prominence-weighted, confirm with Searchable which one you are reading.

What the scaling does not explain

Rescaling Searchable's scores by 1.096 lines up five of the nine brands to within about 2 points of Profound. Four do not move into place:

The two tools normalise brand names differently. Profound labels these companies "Google," "Kling" and "Luma." Searchable labels the same three "Google Veo," "Kling AI" and "Luma Dream Machine." Google and Luma are the clearest cases: one tool names the company, the other names the product. Kling and Kling AI are two forms of the same product name rather than parent and product, so the split is about naming convention generally, not a single consistent rule.

Neither tool exposes its mapping rules, so we will not claim this causes the residuals. It is worth noting that two of the four outliers, Google Veo and Kling, are among the brands where the two tools disagree on what to call the entity.

The two also draw from different entity pools. Profound's Average Position leaderboard includes assets like Frameloop and YouTube alongside the video generators, while Searchable's list is a tighter competitor set. Those are different interface populations rather than a like-for-like measure of detection breadth.

Entity resolution is the least examined variable in AI visibility tooling. It is in neither vendor's marketing, rarely in comparison articles, and it changes who appears in your competitive set.

The practical takeaway

Report rank position and Share of Voice to clients. Both held up when we moved between tools.

Treat Visibility Score as internal to whichever platform produced it. Switch tools and that number can shift by several points on its own, while the brand's actual standing in AI answers has not moved at all. Nobody wants to explain a drop that only happened in the dashboard.

One Visibility Score Hides a Four-Fold Gap Between Engines

Both platforms report a single headline visibility number for each brand. Neither warns you that the number is an average of wildly different results per engine. In our data one brand ran 82% in ChatGPT against 19.1% in Perplexity. That is a four-fold spread.

Profound's Matrix View for our five-day window:

Pika's aggregate Visibility Score of 41.4% compresses 82% in ChatGPT, 22.9% in Google AI Overviews and 19.1% in Perplexity into one number. InVideo runs the opposite way: 26%, 47.9%, 72.3%.

Any brand reading only its headline score is reading an average of three very different situations, which is the strongest argument in this comparison for Profound's Matrix View. It puts competitors in rows and engines in columns and makes the spread visible in one screen. We did not find the same competitor-by-engine matrix in Searchable. The closest view in our workspace was its Topic Position Rankings heatmap, which maps topics against engines rather than competitors against engines. Whichever tool you pick, segment the number before it reaches a client report.

Prompt Research and Query Fanouts: Which Tool Shows More of the Machine's Work

Two questions sit behind every tracked prompt: what you should be tracking, and what the engine actually searches once that prompt lands. Searchable is stronger on the first. Profound shows more of the second. Neither view is complete, and both tools admit it.

What to track

Searchable puts more of this in the self-serve product. Its Prompt Research view showed topic, Visibility Score, average position, sentiment, volume, difficulty, mentions, citation share, and both active and proposed prompts, with generation built in.

Profound has the more distinctive dataset. It describes its Prompt Research Reports as built on a corpus of more than 1.5 billion real user prompts, put through semantic retrieval, filtering, clustering, and intent classification.

But check the plan table before you buy on that basis. Profound's own pricing lists Prompt Volumes searches as None on both Starter and Growth, with only relevant keywords included. Full Prompt Research is Enterprise. If the real-user prompt dataset is why you are choosing Profound, the $399 tier does not include it.

What the engine searches

Searchable showed higher raw fanout counts in the captured screens. 109 distinct fanout queries and 128 total occurrences across the prompt set, ranging from 13 to 23 per prompt, against query counts of 11 to 20 in Profound. Because the captured windows and the counting definitions were not normalised between the two tools, we do not treat that as a volume benchmark.

Profound did more with the fanouts it captured. Its screen separates query count from unique fanout variations, 8 to 17 per prompt, and adds average queries per execution, query variation frequency, word-level transformations showing what the engine adds and drops, and a per-engine breakdown. In our window ChatGPT averaged 2.3 queries per execution against Perplexity's 1.0, which is a genuinely useful diagnostic: one engine explored the category more than twice as widely as the other.

Both carry caveats. Profound's fanout page shows a banner stating that query fanout data is not guaranteed and that coverage varies, with some runs missing it entirely. Searchable's fanout tracking coverage column read 0% on every prompt in our account. Neither view is a complete picture of what the engines actually searched.

Citations and Sources: Which One Produces a List You Can Act On

Counting citations is table stakes now. What separates these two is the list you walk away with: Profound hands you specific pages worth pitching, Searchable hands you the content formats that keep winning. Both are useful. Which one you want depends on your workflow.

Where they agreed

We compared the cited pages each tool surfaced for the same prompt set. Three of the top five overlapped, rising to four of the top six, with the same page in first place in both:

The pattern from the visibility comparison repeats a layer down. The two tools order the list differently but draw from the same pool, so either one will point your outreach at broadly the same pages.

Where they differ

Profound is better at telling you which page to target. Its Top Citation Pages view combines citation share, first-cited date, category, and a Mentioned on Page flag. That flag is the single most actionable column in either product. In our data, zapier.com's roundup was the most-cited page in the category at 4.79% citation share, and it flagged Not Mentioned for the brand tracked in that workspace. That is a specific, prioritised outreach target rather than a report.

Searchable is better at telling you what kind of page keeps winning. Its Sources view classifies every URL by source type (Editorial, Competitor, Review, UGC) and content type across 14 categories. In our dataset, ranked lists accounted for 45.1% of cited content and comparison pages another 21.0%. Two formats, two thirds of the citations. That is a content brief.

The classification gap is measurable. Profound sorted its citation set into Earned Media 19.7%, Social 16.4%, Institution 0.09%, and Other 63.8%. Nearly two thirds land in a bucket that tells you nothing. Searchable's captured taxonomy was materially more granular across the same period, with 14 content types and four source types, and no comparable catch-all category on screen.

Pika scored 41.4% visibility and 0% citation share across the same prompt set. Sixth in the category by Visibility Score, and not once the source. Visibility and citation are separate problems, which is the entire reason citation analysis exists as its own product surface.

Neither tool gets you there on its own. Being recommended in AI answers, not just mentioned in passing, is what GEO optimization is for: showing up in the sources these engines pull from, and giving them a reason to name you when someone asks who to use.

Competitive Intelligence Past the Leaderboard: Matrix View vs Brand Facts

Every AI visibility tool ships a competitor leaderboard, and a leaderboard on its own tells you almost nothing useful. The real question is what each product lets you do once you have that ranking. One slices it apart. The other explains what it means.

Profound slices the ranking across more dimensions. Matrix View by competitor and engine, a prompt-by-prompt top-10 brand matrix, Visibility Rankings by Topic, plus Regions and Personas. The prompt-level matrix is the standout: it shows the top ten brands for each individual prompt, which surfaces the cases where an aggregate leader is actually weak on the specific query you care about. In our set, CapCut ranked first for the TikTok prompt while sitting eighth overall.

Searchable interprets it. Position Distribution breaks out how often each brand lands at #1, #2, #3 and below. Decision Journey maps mention rate to funnel stage. Head-to-head comparison puts two brands side by side with what each is known for. Brand Attributes shows which qualities AI associates with you and at what weight. Brand Facts surfaces specific claims AI platforms make about your brand with correct and incorrect buttons, which is a genuinely novel feature for catching hallucinated product claims.

One practical difference worth flagging: the two workspaces ended up with very different topic structures from the same 10 prompts, 1 topic in Profound against 7 in Searchable. Topic-level analysis is only as good as the taxonomy underneath it, so check what yours looks like before reading anything into a topic breakdown.

AI Crawlers and Referral Traffic: Which Tool Connects Visibility to Actual Visits

Visibility is only the first link in a chain. A brand can be mentioned constantly without being cited, cited without being crawled, and crawled without ever earning a visit. Both platforms now measure past that first link. Setup requirements decide which one you can use.

Profound's Agent Analytics is infrastructure-first. The setup screen supports Amazon CloudFront, Akamai, Fastly, Vercel, Google Cloud CDN, Cloudflare via Worker or Logpush, Netlify, WordPress, Shopify, and Adobe Experience Manager, with Webflow, Squarespace, Framer, and HubSpot listed as not yet supported. Domains are unlimited on Growth. Some integrations carry a dependency worth knowing about: Cloudflare Logpush requires a Cloudflare Enterprise plan, and Netlify requires Netlify Enterprise.

Searchable's LLM Analytics folds crawler and referral data into the same analytics model as citations and visibility, exposing crawler activity by page, referral sessions, top referral pages, sitemap crawl coverage, and crawl-to-visit correlation. Server-side collection is supported through Cloudflare and other infrastructure rather than browser JavaScript alone.

If your stack is on the not-yet-supported list, or you are on standard-tier Cloudflare, check the setup requirements before you buy either one.

The 41.4% visibility and 0% citation share noted earlier is the first link in that chain breaking. Crawler analytics is how you find out whether the later ones hold.

Profound Agents vs Searchable Agent: From Visibility Gap to Shipped Fix

This is where the two products stop competing on the same axis. Profound treats automation as something you build yourself. Searchable treats it as something already running in the background. Neither approach is wrong. They suit very different teams and very different workloads.

Profound treats Agents as a build surface. Templates organised by content creation, content optimization, research, reporting, and outreach. Folders, scheduling, Knowledge Bases, Integrations, Sheets for running agents at scale, and a natural-language builder for custom workflows. Named templates in our account included AEO-Optimized FAQ Generator, Research-backed PDP Optimizer, Multi Language Article Translator, and YouTube Metadata Optimizer. Growth includes 400 credits a month, consumed per run based on complexity.

The shape is: data, configurable workflow, reusable scheduled action. Best when you have a process you want to build once and run weekly.

Searchable treats the Agent as an operating layer. It sits alongside Opportunities, which surfaced 15 items in our workspace without being asked, plus Content, Reports, Site Health, Issues, and a Knowledge Base. The report builder is worth calling out separately: it has 20-plus draggable block types (Visibility Trend, Mentions vs Competitors, Query Coverage Gaps, Decision Funnel, Position Distribution, Brand Perception Treemap) with an assistant that will review, tighten, or prune the report you have built.

The shape is: data, surfaced opportunity, one-click action. Best when you want the diagnosis handed to you rather than designed.

What Each Platform Actually Costs at the Same Tracking Capacity

Everything above is what the two platforms do differently. This is what they charge for it. Entry price hides most of the story: Profound opens at $99 against Searchable's $125, until you check that $99 buys 50 prompts and ChatGPT only.

Brand plans

At 100 prompts the gap is $399 against $125 for the same three engines. In our test both ran daily and each collected 150 answers over five days.

One caveat on the units: Profound counts 100 prompts across 3 engines, Searchable counts 100 per model. They behaved identically on our 10-prompt set, but that is one configuration, not a guarantee at 100.

Agency plans

The agency models are built on different logic entirely.

Profound charges $99 for the shell plus $399 per client workspace. Searchable bundles client slots into the base price. At five clients that is roughly $2,094 a month against $400, assuming one Searchable project per client. Both sides offer prospecting tooling; we tested neither.

The Verdict: Which Platform Should You Buy

Both platforms now cover the same ground: visibility, Share of Voice, competitor tracking, fanouts, citations, sentiment, crawler analytics, AI-assisted execution. The decision is not about feature coverage. It comes down to economics and which kind of analysis your team will actually use.

Searchable is the stronger self-serve value in the pricing scenarios we compared. At the 100-prompt comparison it costs $1.25 per tracked prompt against Profound's $3.99, and that lower price still includes site auditing, content production, API access, source and content-type classification, brand-perception diagnostics, and a report builder we found no equivalent for in our Profound workspace. On the agency side it bundles client capacity that Profound sells separately.

Profound's premium is easiest to justify where a team needs its specific strengths. The Matrix View for cross-engine competitive analysis, which is the best answer we found to the aggregate-score problem. The Mentioned on Page flag on cited URLs, which turns citation data into an outreach list. Execution-level fanout analysis, if you are optimising at that level. Personas and Regions if your work needs them, though Profound's pricing page lists 1 language and 1 region on both self-serve tiers. And Agent Analytics if your infrastructure is on its supported list.

Be careful with the Prompt Volumes argument. It is Profound's most distinctive dataset and the most common reason people cite for choosing it, but Profound's pricing page lists Prompt Volumes searches as None on Starter and Growth. If that is your reason, you are buying Enterprise, not the $399 plan.

And do not migrate on the strength of a Visibility Score. In our test the two platforms agreed on the competitive landscape at 0.88 correlation while Searchable's headline scores ran 3.5 points lower on average, almost all of it calibration. Ranks and Share of Voice travel between tools. The proprietary score does not.

If you are running one platform and reporting to stakeholders, use its Visibility Score as a longitudinal metric within that system and never as a cross-tool benchmark.

For agencies, the economics decide it. Searchable bundles client slots into the subscription, 5 projects at $250 and 10 at $400, with white-label reports and Looker Studio from the entry tier. Profound charges $99 for the shell plus $399 per client workspace, so five clients runs roughly $2,094 against $400.

Most GEO agency and AEO agency work is a repeatable reporting cycle across a book of clients, and that maths favours Searchable. Profound is easier to justify on a small number of high-value engagements where the analysis is the deliverable.