A 10-hour report built in 20 minutes. That's the claim from Supermetrics' launch announcement for their Claude integration. The number sounds like vendor hyperbole until you consider what it actually replaces: logging into Teads, exporting CSVs, cleaning columns in a spreadsheet, pasting into a deck, and then doing it again next week. The new connector lets you query Teads data directly inside Claude using plain English, which means the analyst bottleneck shifts from data wrangling to decision-making.
For marketing leaders running outstream video and CTV campaigns through Teads, this changes the operational math. Teads reaches 2 billion consumers worldwide across premium publisher inventory, and its InRead format has become a staple for brands that want viewability without the intrusiveness of pre-roll. The problem has always been getting that performance data into a format your CFO can interrogate. Now you can ask Claude to break down video completion rates by creative size, compare viewable impressions across device types, or flag which campaigns are burning budget without moving the needle.
The Mechanics: Three Steps, No Engineering
The Supermetrics connector page walks through the setup: open Supermetrics in Claude, authorize your Teads account, and start asking questions. No code, no API keys to manage, no data warehouse required. The integration uses Supermetrics' MCP (Model Context Protocol), which means Claude pulls live data from your Teads account rather than working from a stale export. When you ask "Which campaigns have the highest video completion rate this quarter?", the answer comes from your actual numbers, not a cached file from last Tuesday.
The available metrics cover the full Teads reporting suite: impressions, clicks, CTR, budget spent, total ad cost, unique viewers, video starts, video completion rate, viewable impressions, viewable rate, and both post-click and post-view conversions. Dimensions include advertiser name, campaign name, line item, creative name, creative type, creative size, device, browser, country, and website domain. That's enough granularity to run a proper channel-mix analysis without leaving the chat window.
Where This Actually Saves Time
The real value isn't the query itself. It's what happens after. According to Supermetrics' workflow guide, the biggest time savings come from going directly from raw data to a slide-ready deck or CMO-ready email in a single conversation. You can ask Claude to pull Teads performance, compare it against Meta or Google Ads data (if you've connected those sources too), and then draft an executive summary with the key takeaways. The analyst who used to spend half a day on that workflow can now do it before the 10 a.m. standup.
Cross-platform comparison is where this gets interesting for media mix decisions. Teads' outstream inventory competes for the same budget as YouTube pre-roll, Meta video, and CTV buys. If you can query all of them in the same conversation, you can answer the question that actually matters: where is the next dollar most efficient? The Supermetrics MCP documentation confirms that the same connection works across Google Ads, Meta Ads, LinkedIn Ads, TikTok Ads, GA4, and dozens of other sources. One interface, one conversation, one answer.
The CFO Question: What's the ROI Model?
Marketing AI integrations have a credibility problem. MarTech's recent analysis of AI workflow ROI identifies three measurement dimensions: time saved, output quality, and revenue lift. The Teads-Claude connector scores cleanly on the first. If your team runs 20 campaigns per quarter and each performance review previously took 4 hours of data wrangling, you're looking at 80 hours per quarter returned to higher-value work. At a fully loaded analyst cost of $75 per hour, that's $6,000 per quarter in recovered capacity.
Output quality is harder to measure but easier to observe. When the analyst can iterate on a question in real time ("Show me the same data but exclude mobile," "Now rank by cost per completed view"), the final recommendation gets sharper. The feedback loop between question and answer shrinks from days to seconds.
Revenue lift requires attribution discipline. If faster Teads analysis leads to faster budget reallocation, and that reallocation improves CAC payback by even 5%, the downstream impact compounds. But you need the measurement infrastructure to prove it. The integration doesn't solve attribution; it just removes the data access friction that used to delay the decision.
What This Doesn't Do
Claude is not a BI tool. Supermetrics is explicit that standardized dashboards are possible but require a hosting solution, or Claude can help you build the queries and logic to power them in your existing BI stack. If you need a live dashboard that refreshes automatically and gets embedded in your CRM, you still need Looker or Tableau or whatever your data team has standardized on.
The integration also doesn't replace your attribution model. Claude can tell you that Campaign X had a 72% video completion rate and Campaign Y had 58%, but it can't tell you which one actually drove pipeline. That requires connecting Teads exposure data to your CRM outcomes, which is a different integration problem entirely.
And there's a governance consideration. Supermetrics' MCP documentation notes that each user authenticates individually via OAuth 2.0, with per-session isolation ensuring users see only the accounts they're authorized to access. That's table stakes for enterprise security, but you'll still want to confirm with your data privacy team before connecting production ad accounts to any AI interface.
The Pilot Checklist
If you're evaluating this for your team, here's a two-week test:
- Connect one Teads account (not your largest) to Claude via Supermetrics
- Run three queries you'd normally build in a spreadsheet: campaign performance summary, creative-level breakdown, and week-over-week trend
- Time the old workflow versus the new one
- Have your analyst draft a one-page executive summary using Claude's output
The success metric isn't "did it work?" It's "did it change how fast we made a decision?" If the answer is yes, you have a business case. If the answer is no, you've lost two weeks and a free trial.
The broader signal here is that the data access layer for marketing is collapsing. The moat used to be knowing how to pull the data. Now the moat is knowing what to do with it. Teads-to-Claude is one connector among dozens, but it's a useful test case for whether your team is ready to operate in a world where the analyst's job is asking better questions, not building better spreadsheets.
