Triple Whale

Generally trusted as a measurement layer, though some users voice broader skepticism about whether any attribution can be fully accurate post-iOS 14.

“alright i'm about to drop some sauce on Youtube Ads i'm not here to boast about a measly $9,000/day in spend for a client on youtube ads but, i get questions all the time for understanding how to recognize how to scale youtube ads, how hard to scale then, what do you do when they break, etc. and i'm going to tell you exactly what you need to know, and what 99% of people don't realize about youtube ads this is VERY simple - when you are launching youtube ads, if you are doing high creative volume tests, then you're going to need to group creatives together in campaigns to reduce the overall # of live campaigns for easier management - if you are doing low volume tests, then single ad campaigns work the best - WHEN YOU LAUNCH THE CAMPAIGN, make sure to not immediately set a tCPA or a tROAS, keep it on Max Conversions - if the ad is a winner, we are often seeing that the performance is stable or improving around the target CAC or NC ROAS over the course of 1-2 weeks (depending on budget) - we will THEN increase our budgets. campaigns are much more volatile at lower spend than they are at higher spend - therefore starting a campaign test at $1000 is a much stronger test than at $100, but we all know that most of us can't test 100 creatives/week at $1000/each campaign. - when increasing the budgets, we find it important to use an MTA (Triplewhale, Northbeam, or if you have something for intraday reporting, then that works too) - at lower scale (below $1000-$2000/day in spend), we are not making massive jumps in spend. We can increase the budget from $100 --> $200, but then we need to see the campaign stabilize, then $200-->$300/$400 and repeat. - WHILE KEEPING THE CAMPAIGN ON MAX CONVERSIONS - once you pass $1,000-$2000/day in spend on Max Conv. and maintain or improve your target ROAS or CAC goals, then we can start aggressively scaling. - when trying to aggressively scale campaigns from $2000/day to $50,000/day, what's most important is having a proper way to measure intraday results - we often use TW or NB, and we will watch the campaign nearly every 30 minutes to ensure that the ROAS is holding on both Click Based and Event Based Attribution - as soon as the ROAS starts breaking, that doesn't automatically mean that you decrease the budgets... you should also be spreadsheeting your hourly Overall Spend, Revenue, and MER to see if you're actually noticing a decrease in MER as the ROAS breaks in the MTA - once you can confirm this, we will often decrease the spend by 10-20% and hold the budget there for the next day and monitor the results - the goal is to keep the campaign on max conversions as long as possible, because as soon as you put on a tCPA, you're limiting the further scale of the campaign - this is how we're getting accounts to $30-$50k/day consistently in spend on Youtube time and time again it's a tricky art”— @Nate_Google_ on X

Triple Whale's AI-first pivot draws praise, but its bundled feature push (vs. focused competitors) also draws pushback.

“Can someone tell Triple Whale that all we want is a direct integration with Claude and that's it? Enough of the Moby stuff and this new automation platform, please Integrating with Claude makes your tool 5x more useful. Not integrating makes it 5x less useful”— @dtcprophet on X

Dashboards are a clear upgrade from manual spreadsheet reporting, though some users find specific calculations like profit margins unreliable.

“We’ve come a long way since I was cold-calling customers in a WeWork. Triple Whale was never really supposed to be used outside of the brand I was running. I just wanted to see ad metrics on my phone in real time. “Back in the day, you had to export from the Meta ads manager, and then trying to join it together with customer data, it was nuts.” We saw these problems as jobs to be done and built the tools to execute them.”— @aryehMaxx on X

Ad platform integrations are well liked, but users are vocal about wanting a direct Claude integration instead of more native automation.

“Running ads sometimes felt like day trading with my family’s income. I remember waking up, opening the account, and instantly feeling my mood move with the numbers. If ROAS was up, I felt amazing. If ROAS was down, I felt the pressure immediately. Because it wasn’t just a metric. That ad account was how I paid bills, supported my team, and took care of my family. So anything that could help me operate faster, make better decisions, and remove some of that daily pressure, I wanted it. Media buying is hundreds of small decisions every week: Launch new creative. Move budget. Pause what’s not working. Scale what is. And every decision needed context. What worked before? How aggressive should we be? That’s why we’re releasing Moby Actions today, and why it’s such a big deal. Moby sits on top of your Triple Whale data, so it understands your brand, history, performance, rules, skills, and how you operate your ad account. No MCPs to set up. No APIs to wire together. No custom data workflows to maintain. It runs on the normalized data warehouse Triple Whale already built for you, with your permissions, clean data layer, history, and business context already in place. Then Moby recommends the next best actions: Adjust budget. Launch a creative test. Pause what’s not working. Scale what is. You approve what you want, and Moby executes directly into the ad channels. With Moby’s Harness, you can also compare different models, see how their recommendations differ, and understand why. So you’re not getting a black-box answer. You can evaluate the recommendation, choose the path you trust, and take action. This is the thing I would have been chasing when I was running Madison Braids. Not another dashboard to check. A way to operate better, faster, and with less of that daily pressure on your shoulders. https://t.co/IbCPDqchE1”— @aryehMaxx on X

Moby is the standout feature, with users calling it a proactive agent that rivals a full analyst team.

“🚨HUGE NEWS: Shopify jointly invested $25m in this AI Startup called ‘Triple Whale’ And their agent ‘Moby’ can do the work of 3 analysts in 78% less time I ended up testing it myself for a week, and here are the results: https://t.co/qcnSw4x5v8”— @hasantoxr on X

Value perception splits sharply — evangelists call it essential, while others are frustrated by upgrade friction and subscription cost.

“@triplewhale @AY_Orbach Will you please just let me upgrade my TW subscription without having to book a call?”— @jforjacob on X

Reliability is mixed: strong praise for shipping speed and AI accuracy claims, offset by complaints about calculation errors and tracking discrepancies.

“Attribution is fake Pre-iOS 14, we trusted platform numbers completely. Meta reported 8x ROAS, nobody asked questions. Then iOS 14 hit and most found out "tracking" was largely modeled estimates with attribution windows generous enough to claim credit for purchases made in a different browser tab. Triple Whale followed Third-party tools promised the real numbers They kind of delivered for about two years Then Meta improved their own modeling and suddenly most brands were paying $1K+/month for a dashboard that confirmed what blended ROAS already showed. Now incrementality is the kool aid. And it has real problems nobody talks about. Seasonal bias is brutal. A test run in November versus January gives you different results because underlying demand is different, not because your ads got better or worse. Testing windows are too short for high-AOV products. If your customer thinks about a $1,000 purchase for six weeks, a three-week holdout test captures almost nothing useful. And none of it touches referrals, repeat behavior, or long-term LTV. The math also doesn't work well below ~$1M/month in spend. These platforms run $10K+/month. You end up paying more for the measurement than the insights are worth. What I've seen actually matter: blended metrics, profit trends over time, and testing big structural changes rather than micro-optimizations. A sharp operator with solid unit economics and basic profit tracking will outperform a brand running $120K/year in attribution software that's optimizing toward the wrong signal and not sending it to Meta”— @Peter_Quadrel on X

Support is the weakest link, with slow response times and lack of live help cited repeatedly, especially versus Northbeam.

“why no matter what I do can I not get someone from Triplewhale to hop on a meeting with me to discuss tracking issues? We have brands doing $5m/month that are having major tracking discrepancies and Triplewhale support is not offering time for a call to diagnose this Insane”— @Nate_Google_ on X