docs.clickmax.io

clickmax-vturb

Use when the user asks about VSL / Vturb video performance — play rate, engagement, retention, A/B test winners, whether the video is converting, how many people are watching now, or how to connect the Vturb account.

First seen Aug 21, 2026

Installation

$ npx skills add https://docs.clickmax.io

Also in this package

Other skills from docs.clickmax.io · top by installs.

npx skills add https://docs.clickmax.io

Browse all from docs.clickmax.io

More details

Agent compatibility

Declared targets from SKILL.md / docs. Unmarked agents are not listed — the skill may still install via the CLI.

Claude Code Not declared
Cursor Not declared
Codex Not declared
GitHub Copilot Not declared
Windsurf Not declared
Gemini CLI Not declared
Cline Not declared
OpenCode Not declared

Skill metadata

Parsed from SKILL.md frontmatter.

Versionv12.1.3
LicenseMIT
CompatibilityRequer o MCP da Clickmax ativo e conectado. Os nomes de tool
More metadata
author
Bilhon Technologies LTDA
version
v12.1.3
tools
["vturb_status","vturb_players_list","vturb_player_stats","vturb_player_live_users","vturb_ab_tests_list","vturb_ab_test_stats"]

Package contents

Files included with this skill beyond the listing page.

  • skill md SKILL.md 7,489 B

History

  1. First seen on skills.sh
  2. First recorded snapshot · 30 installs

SKILL.md

As tools abaixo aparecem com os nomes que o MCP da Clickmax registra. Se o seu cliente de IA prefixar nomes de tool (mcp<servidor>, mcp<servidor>, ou outro), use o nome já prefixado que aparecer na sua lista de tools.

When this applies

Questions about the performance of a VSL hosted on Vturb (the video player Clickmax integrates with): plays, play rate, engagement, finishes, clicks, conversions tracked by the player, and near-real-time audience. Also the "how do I connect Vturb" question.

Not this skill

  • revenue/sales totals, top products, funnel steps -> clickmax-analytics
  • payment/transaction cohorts -> clickmax-payments-dashboard-analysis
  • building or editing the page that hosts the video -> clickmax-funnels (page building)

Key assumptions

  • Workspace-scoped and read-only. There is no tool to connect, disconnect, or change the Vturb account — that happens on the Integrations page (see below).
  • One Vturb account per workspace.
  • vturb_status is the mandatory first call — every metric tool returns a precondition error when no account is connected.
  • playerId only comes from vturbplayerslist. Never guess or accept an id from prose.
  • Rates are 0..1 fractions (0.42 = 42%). Multiply by 100 before showing.
  • vturbplayerstats data is cached ~5 min upstream — it is not real-time.
  • Vturb rate-limits: a 429 means back off, not that the integration is broken.

Connecting the account (no tool for this)

If vturb_status returns connected: false, or status: "invalid" (key revoked at Vturb), do not ask the user for the API key — no tool accepts it, and a key pasted in chat would be exposed in the conversation history.

Tell them to connect on the Integrations page, which is where the connection is managed:

  1. Go to IntegraçõesExplorar integrações.
  2. Open the Vturb card (section Vídeo e VSL).
  3. Paste the key generated at Vturb under Settings > Analytics API and connect.

Optional on that screen: pasting the embed code of any video from the account identifies which Vturb account it is, which is what enables picking videos from the page editor. It can also be done later, in the editor itself.

Shortcut while building a page: the Vturb block in the page editor (group Mídia/Elementos) offers the same connection in its right sidebar. Both places write the same connection — there is only one Vturb account per workspace.

status: "invalid" = the key was revoked or expired at Vturb. Replace it with a fresh key on the Integrations page; disconnecting first is not required.

Thought process

  1. vturb_status — connected? If not, give the connection steps above and stop.
  2. vturbplayerslist — resolve the video the user means into a playerId. If several match the prose, ask which one instead of guessing.
  3. Pick the metric tool:

- performance over a period ("está convertendo?", "como foi essa semana") -> vturbplayerstats with startDate/endDate - who is on the page right now ("tem gente assistindo?") -> vturbplayerliveusers - comparing variants ("qual versão ganhou?") -> vturbabtestslist then vturbabtest_stats

  1. Interpret against duration and pitchTime from step 2 (see below) and answer with numbers, not adjectives.

Reading the metrics

|Field|Meaning|Trap| |-|-|-| |plays|pressed play|not unique people — use uniqueDevices for that| |playRate|plays ÷ page loads|low play rate is a thumbnail/headline problem, not a video problem| |engagementRate|average watch time up to pitchTime|a late pitchTime mechanically lowers it — always read with pitchTime/duration| |finishes|reached the end|| |clicks|clicked the CTA|| |conversions / conversionRate|tracked by Vturb, over plays|can diverge from Clickmax sales — never present them as the same number|

pitchTime is the second where the offer is made. Market reference: a pitch around 60–75% of duration is typical; much later means few viewers ever hear the offer.

For vturbplayerlive_users: it counts people who entered the page in the last minutes (default 60). Say "entraram nos últimos N minutos" — not "estão assistindo agora", because it does not prove they stayed.

Diagnosing a weak VSL

Read the funnel in order and name the first stage that breaks:

page loads -> playRate -> engagementRate (até o pitch) -> clicks -> conversions
  • low playRate -> thumbnail, headline, or the video is below the fold
  • decent playRate but low engagementRate -> the opening minutes lose people
  • good engagementRate but few clicks -> the CTA is weak, or hidden when the pitch happens
  • clicks but few conversions -> the problem moved past the video, to the checkout/offer

A/B tests (comparison groups)

A Vturb A/B test enrolls 2+ players (variants of the same VSL) and splits traffic between them by percentage. vturbabtestslist gives the tests; vturbabteststats compares up to 2 variants per call (upstream limit — for a 3-variant test, compare in pairs).

Declaring a winner:

  • Decide by revenuePerVisitor, not conversionRate. A variant can convert less and still make more money per visitor (higher ticket). This is the single most common misreading.
  • If running is true the test is still splitting traffic — present numbers as a trend, never as a final verdict.
  • Watch the sample size: with few views per variant, a difference of a few points means nothing. Say so instead of crowning a winner.
  • locked on a variant means its traffic share is pinned in the Vturb panel.

Fair window: omit startDate/endDate so each variant uses its own started_at inside the test. Forcing a shared window penalises a variant that joined later. Only pass explicit dates when the user asks for a specific period.

retentionCurve ({ second, viewers }[]) comes with these stats — it is the only place we expose per-second retention today. Use it to locate the drop-off and to choose the progressive-reveal second (below).

Progressive reveal (using the data to change the page)

The Clickmax Vturb block supports "Revelação progressiva": reveal or hide page elements at a given second of the video (data-vturb-timer). When the user asks when to show the CTA, anchor the recommendation on pitchTime from vturbplayerslist — revealing the button around the pitch instead of at page load is the common fix for "engaja mas não clica".

Per-second data is only available for videos enrolled in an A/B test, via retentionCurve in vturbabtest_stats. For any other video do not claim to know the exact drop-off second — reason from pitchTime, engagementRate and finishes.