Author · Senior Data Scientist
Dr. Yang Qi
Senior Data Scientist, SOMIN · 13 articles on the SOMIN blog
About Yang
Dr. Yang Qi is Senior Data Scientist at SOMIN, the Singapore AI marketing-intelligence company. Yang writes on the SOMIN blog about digital advertising, performance advertising, insights.
Profiles
Articles by Dr. Yang Qi
13 articles
AIDA model examples: what does each stage look like in a real campaign?
Seven AIDA model examples across B2B, retail, services and APAC, showing what attention, interest, desire and action look like in an ad, email or page.
What is the 2027 marketing calendar for Korea, from Seollal to Chuseok?
Korea's 2027 marketing calendar: Seollal on 7 February, Chuseok on 15 September, Pepero Day on 11 November, plus holiday breaks and shopping events.
Meta Werbebibliothek: how do you analyse competitor ads in Germany and the EU?
The Meta Werbebibliothek (Ad Library) lets you search a rival's ads in Germany. Ads aimed at the EU are kept a year, with targeting details. Here is how.
What is the Ramadan and Eid al-Fitr 2027 marketing calendar for MENA and Southeast Asia?
Ramadan 2027 is expected to start on 8 February and Eid al-Fitr on 9 or 10 March, depending on the moon. See dates for the UAE, Saudi Arabia and Southeast Asia.
What is shadow AI governance, and what should a marketing team's one-page policy say?
Shadow AI is unapproved AI use at work. Govern it with a one-page marketing policy: approved tools, data classes, review, logging and incident steps.
Why do marketers distrust AI insights, and how do you make them verifiable?
Marketers distrust AI insights because they cannot see where conclusions come from. Make them verifiable by tracing every claim to its sample and evidence.
What is creative clickability, and can you predict it before spend?
Creative clickability is an ad's predicted chance of beating comparable ads on clicks, scored before spend. How the models work, where they fail, how to test.
What gets B2B martech content engagement on LinkedIn?
B2B martech posts win on LinkedIn when a person writes them, they use one image and they show people or events. Findings from 734 vendor posts, Jul–Oct 2026.
Pre-testing creative: how to predict ad performance before you spend
To predict ad performance before you spend, score each creative with a model trained on labelled ads, cut the weak tier and live-test only the close calls.
How to measure social media influence beyond follower counts in 2026
Measure social media influence by engagement per post, shares, saves, retention, topic share and sales, not follower counts. A metrics table and creator cases.
AI content homogenisation: why every brand sounds the same
AI content homogenisation happens when brands give similar models similar briefs. Why every brand starts to sound the same, and how context data fixes it.
How to build market personas from audience data and competitor ads
Build market personas from audience data by pairing panel data with competitor ads in each market, then validating the clusters. Lessons from two SOMIN cases.
How do Google and Meta smart bidding algorithms differ?
Google's Target ROAS bids for conversion value from campaign data; Meta's smart bidding ranks ads on predicted engagement from user data. How they compare.