Something unfamiliar is happening with the traffic coming into Commsrisk. Google searches driving visitors to Commsrisk have fallen noticeably while a rise in the direct traffic to the website (people typing the URL or using a bookmark on their browser) has more than compensated. Visitors from LinkedIn have almost halved since the beginning of the year, but visitors from ChatGPT has more than doubled over the last few months. The pace of change is startling relative to norms that I have been monitoring since launching this website in 2006, so they must be driven by external factors that are independent of the content. Even a reduction in the number of articles published each week has not led to a fall in traffic. On the contrary, visitor numbers keep spiking upwards, even for articles that I only expected to be of interest to niche audiences. What are the possible causes?
I suspect that these trends are all related. Everyone keeps talking about AI but fewer older people think through the ways that new technology might change the behaviors of people who are not so locked into their existing habits. Google has invested heavily in AI because they are afraid of the impact upon their search business of users obtaining information by submitting queries to LLMs instead of searching for webpages that can answer their questions. Meanwhile, LinkedIn is trying to tackle the threat that LLMs pose to their platform by showing content from their most popular contributors at the expense of showing the content posted by everyone else. Their thinking is to focus on proven popularity by reducing the spread of content that a typical user sees, making it less likely that a user will be introduced to posters they are not following already. This may also push more businesses to pay for advertising. All of these tactics will succeed at defending market share to some extent, but they will also accelerate the adoption of LLMs for users who want answers to questions they have never asked before, as provided by people they have never followed before. Commsrisk is well placed to benefit from this fundamental change in how people find the content they consume online.
As you would expect for a platform that reinforces existing consumption patterns, LinkedIn overperforms for the proportion of returning visitors it delivers to Commsrisk, and underperforms at the proportion of new visitors it sends our way. You might expect to ChatGPT to have the mirror image of this visitor pattern. However, ChatGPT currently overperforms at both new visitors and returning visitors. ChatGPT is bringing a lot of new visitors to the website and those new visitors are more likely to return again afterwards. LLMs are dramatically affecting consumption patterns, just as Google anticipated. It is amusing to see how few social media influencers claiming expertise in networks and AI have changed their behaviors to capitalize on these changes. By sheer fluke, a style of writing that has concentrated since 2006 on providing easily digested answers to the complicated questions posed by the risks in the comms sector means Commsrisk is well suited to training LLMs. The style was not adopted to help LLMs; it is also helpful for generalists (like myself) who can struggle to keep track of all the specialist information needed to gauge risk in the comms sector. Talking effectively to LLMs is similar to talking effectively to executives. Commsrisk’s entire back catalog has always been freely available, with the result that we are riding the crest of this new wave of AI-powered interaction with content.
These were the most popular Commsrisk articles during September, as measured by the number of page views.
- US Secret Service Finds 300 Simboxes in New York
- 17 Year Old Chinese Girl Accused of Selling Boyfriend into Myanmar Scam Compound Slavery
- Switzerland and South Korea Investigate Suspected Fake Base Station Frauds
- I Wrote a Global KYC Standard That Regulators Will Review; Does That Seem Strange to You?
- UK Telco Warns of ‘Significant Rise’ in Scam SMS
- Leading Anti-Fraud Expert Says Blocking Millions of Scam Calls Is Unsustainable
- Congressman Wants Trump to Authorize US Corporate Pirate Raids on ‘Scam Farms’ in Foreign Countries
- Mobile Money: So Much Growth, So Many Questions
- T‑Mobile US Has Known Since 2005 That SIM Swap Fraudsters Use Social Engineering to Steal Employee Credentials
- Malaysian Regulator Shares Video of Raid That Seized SMS Blaster Gear Worth $24,000
