ChatGPT’s high-reasoning mode acted for illustration a different hunt aboveground for marque visibility successful a Semrush study pinch Kevin Indig. It cited different domains than successful minimal reasoning mode and ran astir 5 times arsenic galore web searches.
By the numbers. Only 25.6% of cited domains overlapped betwixt minimal and precocious reasoning for the aforesaid prompts.
- Nearly 3 successful 4 sources changed erstwhile ChatGPT shifted from Instant-style answers to Thinking-style answers.
Thinking mode utilized much sources. Citation rates roseate from 50% successful minimal reasoning to 68% successful precocious reasoning. Cited answers besides utilized much sources, rising from 2.6 to 4.5 citations per response. High reasoning ran 1,130 web searches crossed the trial set, versus 245 for minimal reasoning.
Reddit mislaid ground. Reddit’s citation stock fell from 15% to 7% erstwhile precocious reasoning was on. User-generated contented and reappraisal sites dropped from 14.3% to 6%.
- Government and world sources roseate from 1.9% to 8.8%.
- Official archiving and support pages grew from 12.4% to 17.5%.
Comparison prompts drove searches. At the comparison stage, precocious reasoning averaged 24 sub-queries per prompt, versus 5.5 for minimal reasoning. Average citations besides peaked there, astatine 9.8 per high-reasoning consequence versus 5.8 for minimal reasoning.
- For example, a CRM comparison whitethorn trigger abstracted searches for pricing, integrations, security, support pages, and archiving earlier ChatGPT forms its answer.
Early citations lasted longer. High reasoning was much apt to transportation a marque from early investigation into later buying questions.
- In 4 of the 20 journeys tested, a marque cited astatine the problem shape still appeared astatine the action stage. Minimal reasoning showed nary full-journey persistence (whether a marque cited astatine the Problem shape survived to the Selection shape of the aforesaid journey).
- High reasoning besides reused the aforesaid domains much often wrong a azygous answer. The aforesaid domain appeared aggregate times successful 51 of 100 high-reasoning responses, versus 26 of 100 minimal-reasoning responses.
Finance saw the biggest jump. The citation assistance varied by category:
- Finance had the largest increase, pinch citation rates rising 28 percent points successful precocious reasoning.
- Health and manner roseate 24 points.
- B2B SaaS gained 16 points.
- Consumer tech hardly moved, rising only 4 points.
- Even though precocious reasoning ran much sub-queries for user tech prompts than for immoderate different category, it often landed connected the aforesaid brands and sources arsenic minimal reasoning.
Why we care. Your contented whitethorn look successful accelerated ChatGPT answers but vanish erstwhile users inquire much analyzable questions. Visibility depends connected whether your pages, documentation, and third-party references tin aboveground crossed the smaller searches ChatGPT runs earlier it answers.
About the data. Semrush and Indig tested 100 prompts crossed 20 purchaser journeys successful B2B SaaS, finance, user tech, and wellness and lifestyle. Each punctual ran erstwhile successful minimal reasoning and erstwhile successful precocious reasoning. The study tracked citation rate, cited sources, and fan-out queries.
The report. Only 25% of cited sources overlap betwixt ChatGPT’s different reasoning modes [Study]
English (US) · 
Indonesian (ID) ·