Your Industry Already Decided If Blogging Still Works
Your agency answered the wrong question. You asked whether blogging still works, they sent back a traffic chart, and that chart cannot tell you what to do on Monday, because in 2026 the answer stopped being a percentage and turned into a category. Blogging still works in cybersecurity and marketing software, but in the top-of-funnel sense you were sold, it has stopped working across finance, healthcare, legal and consumer tech. If your business sits on the losing side of that line, better articles will not carry you back across it, and every month you keep funding them is the most expensive line on your marketing invoice.
The Famous 42% Drop Is Hiding an Industry Split
Every pitch deck this year quotes the same scary number at you. Define Media Group analyzed organic search clicks across 64 sites and found organic search clicks down 42% since AI Overviews began expanding, reported by Danny Goodwin in Search Engine Land on March 12, 2026. AI Overviews are the AI-written answer block that sits above the blue links. Traffic fell 16% immediately after the launch and never recovered, then kept sliding to 42% below the pre-launch baseline by the fourth quarter of 2025, against a baseline of roughly 1.7 billion clicks a quarter. If half your new customer inquiries used to start with a search, a drop of that size is not a reporting problem, it is a hire you quietly postpone.
That same report carries a second set of numbers that rarely travel with the first. Breaking news traffic grew 103% between November 2024 and early 2026, and Google Discover grew 30%. AI Overviews appeared on about 15% of news queries, roughly three times less often than on health and science queries, per Ahrefs data cited in that reporting. Google is not answering every category at the same rate. It is answering your category at a rate specific to your category, and if you run a clinic, the health and science figure is the one to sit with.
On August 26, 2026, Adam Gnuse published a Search Engine Land analysis of Semrush organic traffic estimates for 30 large publishers across nine industries, comparing August 2026 against August 2025. Every single domain in finance, healthcare, legal and consumer tech lost top-of-funnel traffic year over year. Top of funnel means the early informational articles people read long before they buy anything. Every single domain in cybersecurity and marketing and sales software improved. Cloud infrastructure came out collectively positive. LegalZoom.com fell 90% year over year at the top of the funnel, an outlier attributed to changing Google Search patterns.
Inside every industry studied, top-of-funnel and middle-of-funnel content moved in the same direction. Losing categories lost both. Winning categories won both. That consistency is what makes industry the variable worth arguing about. It is not that some blogs are written better. Whole categories moved together, in the same month, regardless of who was writing.
I read what currently ranks for the question owners actually type, some version of "is blogging still worth it for my business in 2026." Those pages argue the general case beautifully. Not one of them tells you the answer now splits by industry, names which industries lost and which gained, or tells an owner in a losing category what to fund instead. That gap is personal for me. The 300-plus businesses I have worked with over 17 years sit in medical, ecommerce, local services, fintech, education and real estate. Two of those, medical and fintech, overlap with two of the losing four, and the others are not identified in the public data, which tells you how thin the evidence gets once you leave the big categories.
| Industry | Top-of-funnel traffic, Aug 2025 to Aug 2026 | Detail worth knowing |
|---|---|---|
| Finance | Every domain declined | Middle-of-funnel content moved the same direction |
| Healthcare | Every domain declined | Same pattern across top and middle of funnel |
| Legal | Every domain declined | LegalZoom.com fell 90% year over year, a drop attributed to changing Google Search patterns |
| Consumer tech | Every domain declined | Same pattern across top and middle of funnel |
| Cybersecurity | Every domain improved | Gains held across the funnel, not just at the top |
| Marketing and sales software | Every domain improved | Gains held across the funnel, not just at the top |
| Cloud infrastructure | Collectively positive | Reported as a group rather than domain by domain |
Thirty Publishers Are Your Early Warning Light, Not Your Forecast
Before you carry that table into your next marketing meeting, know exactly what it is. The sample is 30 large publishers, not 30 small businesses. A publisher lives on informational traffic, ranks for a huge spread of early-stage queries, and feels a change in Google's answering behavior within weeks. Your two-location physio clinic does not work that way, and nothing in that study promises anything about one clinic's blog.
What publisher traffic gives you is an instrument reading. It shows you, faster than your own analytics will, how Google has decided to answer the questions in your category. Direction, not destiny. I have argued before that the zero-click panic is a publisher problem rather than a small-business one, and I stand by every word of it, because your revenue was never coming from informational pageviews. A legal publisher losing 90% of its top-of-funnel traffic still tells you something true about legal. Ignoring that reading is how owners keep paying full price for a channel that has already been re-priced underneath them.
So check your own category with your own data instead of arguing with somebody else's chart. Five signals do the job, and four of them take under an hour.
Three or more of these and the debate about whether blogging works in general is over for you. It works. It works somewhere else.
Owners Buy More Content to Fix a Verdict Content Cannot Change
The reflex when a blog stops pulling is to buy more blog. Traffic drops, the dashboard turns red, and the obvious diagnosis is quality, so the answer becomes more articles, longer articles, better briefs, a bigger retainer. Nobody in that meeting asks the only question that matters: is this still a question Google sends anyone to a website for?
The spending trend is walking straight into it. Constant Contact's Q1 2026 small business survey, covered by Ayaz Nanji at MarketingProfs, surveyed 1,500 small business owners across Australia, Canada, New Zealand, the UK and the US. Sixty-eight percent expect their marketing budgets to increase in 2026, and 74% expect to spend more time on marketing. Fifty-four percent already use AI marketing tools, with another 27% planning to start. More money, more owner hours, and dramatically cheaper production, all aimed at a channel whose returns now split by category.
Cheaper production is what changed the value of a competent article. When anyone can produce a clean explainer in minutes, a clean explainer is worth roughly what it costs to make. Duane Forrester framed the test more precisely than I have managed to in 17 years of trying.
Run that test across your own archive tonight. The "what is" post, the "how does it work" post, the "10 tips for choosing a provider" post. A summary replaces all three completely, in every industry, without ever mentioning your name. I have watched every "SEO is dead" cycle since 2008, and this one differs in a specific way. The earlier ones changed who ranked. This one changed whether ranking produces a visit at all, for an entire class of question, in an identifiable list of industries.
What Survives Is What a Summary Cannot Reproduce
The blog format itself came out fine, which surprises people who read only the headline numbers. Seer Interactive studied 47,097 citations across ChatGPT, Gemini and Perplexity from March to June 2026, published by Sonny Vasquez on July 24, 2026, and found blogs and guides accounted for 50% of all cited pages. Half of everything AI assistants pointed at was a blog post or a guide. The format is not the casualty. The generic instance of it is.
Adam Gnuse's earlier Search Engine Land analysis, published May 27, 2026, put numbers on which instances survive. Working from Google Analytics data across 10 websites and about 150,000 indexed pages in March 2026, he found educational how-to content earned citations in AI answers roughly 12% of the time. Trends-and-analysis posts earned them 78% of the time. Data-based year-in-review posts, 61%.
| Content type | How often it earned an AI citation | What that means for your budget |
|---|---|---|
| Educational how-to posts | About 12% of the time | The default output of most content retainers, and the least likely to be cited |
| Data-based year-in-review posts | 61% | Needs numbers you own or compile, which is why competitors cannot copy it in an afternoon |
| Trends and analysis posts | 78% | Needs a stated point of view, which a summary has nowhere to borrow from |
Page type matters as much as post type. In the same data, measured as sessions from AI assistants per 1,000 organic sessions, service and product pages pulled 29.4, articles pulled 23.4, FAQ pages 14.0, tool and demo pages 9.8, and homepages 5.6. Your service pages are working harder in AI answers than your blog is. Most owners have never once had those rewritten, while the blog gets touched every week.
Freshness turned out to carry real weight. In the Seer data, 75% of cited pages had been updated within the last year and 88% within two. Pages cited in all four months of the study, the ones with durable visibility, had a median update age of 0.47 years, and 68% had been touched within the year. Pages that spiked for a single month were fresher still, 86% updated within the year, median age 0.16 years.
“Publish and forget loses. Publish and maintain wins.”
Sonny Vasquez, Seer Interactive, July 24, 2026.
That single line rearranges a content budget faster than any keyword tool. Gnuse also points out that interactive tools are one of the few early-stage formats AI summaries will happily cite but will not reproduce themselves, because a calculator that returns your number is not something a paragraph can imitate. His closing test for any page is the one I now use with clients before a single word gets commissioned.
“Does this page deserve to be published?”
Adam Gnuse, Search Engine Land, August 26, 2026.
Most content calendars lose a large share of their entries to that question, and watching an owner run it against their own list is always the moment the retainer conversation changes. The survivors are always the same shapes: your own data, your named people saying something only they can say, tools that do a job, and pages that answer a buying question rather than a curiosity question. If your calendar is full and your pipeline is not, the problem is upstream of the writing, in how the topics get chosen in the first place.
Move the Budget Line by Line, Starting This Week
Nothing about this requires an agency change or a new platform. It requires deciding which line items stop and which start, and doing it before you approve the next invoice.
Step one you do yourself, in an afternoon, with a spreadsheet and no software. Export your published URLs, sit with the list, and be honest about each one. Owners almost always guess low, then find far more of the list is replaceable than they expected.
Step two is a conversation, not a project, and the wording matters. Do not ask your agency whether blogging still works, because you will get a chart. Ask which specific articles they wrote that an AI answer replaces entirely. A good partner will name them without flinching and propose something better. A weak one will explain that content is a long game, which is how most SEO budgets end up still fighting the last war two years after the war moved.
Steps three and four are the only genuine defense in a losing category. Your own numbers cannot be summarized out of existence because they do not exist anywhere else to summarize. A clinic that publishes its real average wait time, a lender that publishes its real approval rate by profile, a physio that publishes a few hundred anonymized recovery timelines. A summary can borrow your topic. It cannot borrow a number only you hold. Delegate the building, keep the numbers under your own control, and never let a writer estimate one.
Step five costs nothing and gets skipped anyway. Three quarters of cited pages in the Seer study had been updated within the year. Ten refreshed pages will do more for you next quarter than ten new ones. A refresh means updating the numbers, adding what changed this year, cutting what is no longer true, and changing the visible date, an afternoon per page.
Traffic Was Never the Scoreboard. A Booked Calendar Is.
Whatever you cancel this month, change the report before you change the spend, because the old report will argue for the old spend every time.
The mismatch runs both ways. The top 10 organic pages in that data captured 55% of organic sessions but only 29% of sessions arriving from AI assistants, and 14% of the pages receiving AI traffic had zero organic clicks in the same window. Pages your rankings dashboard treats as failures are quietly bringing you people. Those people also behave differently: average engagement time was near identical overall, 46.9 seconds organic against 47.1 from AI, but tool and demo pages held AI visitors for 146 seconds against 101 organic, and homepages held them 82 seconds against 36. Someone arrives already convinced, because an assistant vouched for you before they clicked.
The vanity traps are the ones that survive because they are easy to export. Total sessions, keyword count, number of ranking terms, impressions, publishing cadence. Every one of those can rise while your booked calendar empties. Replace them with four: booked calls and qualified inquiries by month, sessions from AI assistants tracked as their own channel (ask whoever manages your analytics to group visits from ChatGPT, Perplexity and Gemini so you can see them in one line), how often you appear in the AI answers for your ten most valuable customer questions (type each one into ChatGPT and Google once a month and record yes or no), and how many of your top pages were updated in the last 90 days. Review that quarterly, not weekly. Category-level shifts take months to show, and watching them daily produces panic decisions.
A London ADHD clinic I worked with ran booked solid for three straight months. They hired more specialists and outsourced the overflow. A full calendar forced a hiring decision. That is what a real scoreboard does: it changes what you do next, not how you feel about a chart.
Frequently Asked Questions
Is blogging still worth it for a medical clinic in 2026?
Not in the form most clinics are sold. Healthcare was one of four industries where every publisher studied by Search Engine Land in August 2026 lost top-of-funnel traffic year over year, and health and science queries draw AI Overviews about three times as often as news queries do. General symptom and condition explainers now get answered on the results page without a click. What still earns its budget is content only your clinic can produce: your own outcome data, your named clinicians answering questions in their own words, your real wait times and pricing, and service pages built to make booking easy.
How do I know if my industry is one of the losing categories?
Start with the four named in the August 2026 Search Engine Land analysis: finance, healthcare, legal and consumer tech. If you are not in one of those, run the test yourself in an hour. Open Search Console, compare the last twelve months against the prior twelve, and check whether your rankings held while your clicks fell, because steady positions with falling clicks is a category signal rather than a quality problem. Then type your five most common customer questions into Google and count how many are answered completely before any link appears.
Should I stop blogging and move the money into ads?
Stop the replaceable blogging, not all of it. Blogs and guides still accounted for half of all cited pages in Seer Interactive's July 2026 study of 47,097 AI citations, so the format is not the casualty. The generic version of it is. Move the canceled budget into work a summary cannot reproduce: your own data, named expertise, one interactive tool, and the service pages that turn interest into a booking. Ads deserve funding for the same reason: they buy the bottom of the funnel that used to arrive free through the top.
The question was never whether blogging works. It was whether a search engine still needs to send a person to your website to answer a question in your category, and in finance, healthcare, legal and consumer tech it has mostly stopped needing to. If you want to know which side of the line your category sits on and what your budget should fund instead, book a call and we will go through your own numbers together. One thing to check before you approve the next content invoice: if a competitor in your category stopped publishing a year ago and their bookings held steady, the articles were never the thing that was working.
