June updates from Swarmgram:
https://t.co/Ll8fDRgCMn - a/b testing live, free sandbox refined.
https://t.co/iqnTv7ZX5g - free chatbot with near instant results. Poll 1000 people in seconds
https://t.co/E3GSqqaoqa - support us by buying a t-shirt or hat!
Virginia is LIVE on https://t.co/VFA11XZRtX with 20,000 demographically calibrated agents! 170,000 agents in total across 11 live markets.
If you're looking to run a study, test marketing materials, or get feedback on your website, comment below for a free study.
STOP WASTING MONEY ON ADS BEFORE YOU TEST!
Run A/B testing on your document, materials, or URL and get up to 10,000 synthetic respondents to tell you what's wrong with your copy FOR FREE.
Try now on Lewsearch; ๐ in bio.
Run A/B creative testing, ad copy testing, or have up to 10,000 census calibrated agents give feedback on any website.
N=200 standard panels are free. Full studies start at $29.
Run n=200 studies for free on Lewsearch. Link in bio.
Lewsearch studies have an average 7% MAE (mean absolute error) compared to ground truth human panels, getting lower every day.
Full studies with 10-15 page PDF's start at $29 each and are between n=500 and n=10,000.
Will be releasing Lewsearch next week to the public.
Soon users will be able to run n=200 studies for free within a few percentage points of ground truth.
@claudeai How are limits? We are spending $400/day on cursor . Do you allow additional spend past max plan, and do you allow using other models occasionally such as codex?
Some sample benchmarks from tonight. Extremely impressive. Average 8.6% MAE on 120+ benchmark questions dragged up by outliers. Full methodology this week.