# Measured vs Recast: which do AI models recommend for Attribution & MMM, October 2026

GTM AI Recommendation Index, October 2026 Edition, Attribution and marketing mix modeling. Zero of fourteen models named Measured first on the direct prompt; zero named Recast. Page: https://gtm-ai-index.com/marketing/attribution-mmm/measured-vs-recast/

| | First-choice share | Rank | Negative rate | Labels | Models naming it |
|---|---|---|---|---|---|
| Measured | 5% | #6 of 17 | 5% | 21 | 13 of 14 |
| Recast | 3% | #7 of 17 | 13% | 15 | 9 of 14 |

## The direct prompt, model by model

- GPT-5.4 mini: neither first, one named (first choices: Dreamdata) (alternatives: HockeyStack, Measured)
- DeepSeek V4 Flash: neither first, one named (first choices: Dreamdata, Rockerbox) (alternatives: HockeyStack, Marketing Evolution, Measured)
- Kimi K2: neither first, one named (first choices: Dreamdata) (alternatives: Factors.ai, Recast, Rockerbox)
- Claude Haiku 4.5: neither named (first choices: Dreamdata, HockeyStack) (alternatives: CaliberMind, HubSpot Marketing Hub, Rockerbox, SegmentStream)
- Gemini 3.5 Flash: neither named (first choices: CaliberMind) (alternatives: Dreamdata, HockeyStack, RevSure, SegmentStream)
- Perplexity Sonar: neither named (first choices: Dreamdata) (alternatives: HockeyStack, Improvado, SegmentStream)
- Grok 4.1 Fast: neither named (first choices: Dreamdata) (alternatives: Factors.ai, HockeyStack, Rockerbox)
- Mistral Small: neither named (first choices: Dreamdata, MixModeler) (alternatives: Northbeam)
- Llama 4 Maverick: neither named (first choices: Rockerbox) (alternatives: Attribution, Dreamdata, Northbeam)
- Qwen 3.7 Flash: neither named (first choices: Rockerbox) (alternatives: Dreamdata)
- GLM 4.7 FlashX: neither named (first choices: Dreamdata) (alternatives: Cometly, HubSpot Marketing Hub, Keen Decision Systems, Rockerbox, Ruler Analytics)
- MiniMax M2.5: neither named (first choices: HockeyStack, Marketing Evolution) (alternatives: Adobe Marketo Measure, Dreamdata)
- GPT-6 Luna: neither named (first choices: HockeyStack) (alternatives: Dreamdata, Paramark)
- Muse Glimmer 30B: neither named (first choices: Dreamdata) (alternatives: Admetrics, Adobe Marketo Measure, HockeyStack, Rockerbox)

## What the models said about Measured

- "platforms (like Recast, Measured, or Rockerbox) routinely cost upwards of $50,000 to $150,000+ per year" (Gemini 3.5 Flash, budget prompt, soft negative)
- "If you want a more turn-key, CFO-ready dashboard with heavy, automated calibration out-of-the-box, look at Measured or Rockerbox." (Gemini 3.5 Flash, scale prompt, first choice)
- "Measured ranks first because it turns MMM from a quarterly report into a weekly operating system." (Muse Glimmer 30B, comparative prompt, first choice)
- "Measured specializes in marketing mix modeling for mid-market companies" (GLM 4.7 FlashX, paraphrase prompt, first choice)

## What the models said about Recast

- "Enterprise-grade, managed MMM and multi-touch attribution (MTA) platforms (like Recast, Measured, or Rockerbox) routinely cost upwards of $50,000 to $150,000+ per year." (Gemini 3.5 Flash, budget prompt, soft negative)
- "Recast has limitations if you need custom transformations—like modeling promo mechanics with different lag structures per SKU." (Claude Haiku 4.5, negative prompt, soft negative)
- "Recast | Self-serve Bayesian MMM, fast refresh, no data science team needed | Companies wanting modern MMM without heavy lift" (Kimi K2, scale prompt, first choice)
- "If you have 1-2 analysts with intermediate statistics experience, look closely at Recast or Haus." (Gemini 3.5 Flash, scale prompt, first choice)
- "Add MMM with: Recast ... offers the best balance of modern methodology and mid-market accessibility" (Kimi K2, paraphrase prompt, alternative)

Share is the count of first choices across the direct, paraphrase, budget and scale prompts over all fourteen models, for a mid-market B2B company; rank is within the category. Comparisons are drawn for the top eight products in each category. Published under CC BY 4.0; the output is the models' output, and nothing here is a recommendation by the index.