Feedly UX Teardown: The Homepage Stopped Talking to Its Own Users

Feedly is the company that caught the users Google Reader dropped in 2013. Type feedly.com into a browser in 2026 and the first line you get is "The fastest way to track emerging threats with AI." The second names the audience: CTI teams. Cyber threat intelligence.

The reader still exists. It still works. It is reached by a button in the top-right corner that says "Go to News Reader." That button connects what most people think Feedly is to what the site now sells, and it is doing that job alone.

The homepage belongs to a different reader

Read it as a stranger and it's a competent enterprise page. The trust line is institutional — "Trusted by 380 CTI teams, including half of the Fortune 10" — and the testimonials at the top are attributed by role rather than name: a CTI Lead at a Fortune 100 financial services firm, a Head of Threat Research at a global pharma company. That anonymity is a deliberate signal. In security, the buyer who can't say publicly what they use is the buyer worth having.

Then comes a wall of counted things: 10,000+ sources, 1,000+ AI models, 10M+ articles, 979 threat actors, 300K+ CVEs, 12K+ malware families, 800 TTPs, 681M+ IoCs. Feedly used to count feeds. Now it counts adversaries.

The page closes on a 30-day evaluation plan broken into Day 1, Week 2–3, Week 4 — a procurement artifact, laid out as a calendar. On the whole page, the word feed survives in exactly one place: the company's own name.

Here are the three pages a reader can still reach, in the order they'd hit them:

The reader is intact — and it's a sub-site

Look at the top-left. The wordmark is followed by a divider and the words "News Reader," with its own Pricing link and its own Resources menu. That divider is the whole architecture in one glyph. This is a neighbouring site.

The page itself is well preserved. "Take control of your news feed." One CTA, "Get started for free," with "No credit card required" underneath. A product shot of the actual reader, ringed by the logos of sources you'd follow — the BBC, the FT, TechCrunch, PlayStation. Scroll and you get "Loved by millions of readers," with quotes from Jason Hirschhorn of REDEF, Brad Feld of Foundry Group, and Tina Roth Eisenberg of Swissmiss. Then Boards, search, mobile apps, and a closing section titled "Privacy by default."

Two things stand out. The first is that those names date the page: they are the blogosphere's cast, circa the years when losing Google Reader felt like a civic event. The second is that across the whole page, AI is never mentioned — on a site whose homepage can barely finish a sentence without it.

Preserved is not the same as maintained. Nothing here is broken. Nothing here has moved either.

Discover

The Discover directory is the strongest argument the reader still has. A search field that accepts a topic, a website, or a raw RSS link. Four tabs — Websites, Reddit, Newsletters, Google News — acknowledging that the open web now reaches people through several different pipes. Then a taxonomy laid out like a magazine rack: Industries, Skills, Fun. Advertising, biopharma, energy, telecom. SEO, photography, data science, writing. Comics, gaming, food, dating.

Every category carries one named publication as its ambassador: Adweek, The Hacker News, xkcd, smitten kitchen. Nobody can answer "what should I follow about biopharma?" from memory. Everyone can react to Science Magazine. That's the recognition principle, applied to the hardest problem a feed reader has (NN/g).

The hardest problem is the empty state, and this page solves it with editorial judgement rather than an algorithm. A human decided that smitten kitchen is what Food means. A human also put Cybersecurity in the four featured cards, on the reader's own directory page.

The changelog is the only page that tells you the truth

Marketing pages are written. Changelogs accumulate. That makes this the most reliable document on the site.

The nav lists the company's three products in priority order — Threat Intelligence, Market Intelligence, News Reader. The filter row below offers All, News Reader, Fundamentals, Market Intelligence, Threat Intelligence. And then the entries: enriching Atlassian and Apple security releases in the Vulnerability Intel Agent, VirusTotal verdicts, GreyNoise context on IoC cards, an OpenCTI knowledge-graph sync, SOC 2 Type 2, an MCP server for automating CTI workflows with Claude.

The entire first page of entries is Threat Intelligence, Market Intelligence and one navigation improvement. The News Reader filter is honest infrastructure attached to a product that has stopped generating rows.

What the seam costs

Leaving the reader exactly as it is may well be the kindest available option. It works, it's private, and nobody is being asked to migrate.

The proof has stopped keeping time. Testimonials are a liveness signal. A returning reader lands on a homepage addressed to somebody else, clicks through to their own product, and finds its newest social proof is a decade old. By the third stale name they have concluded the reader is finished, and the company will never get to argue otherwise.

Discover is the bigger loss. It's a curated map of the open web, maintained by someone with taste, filed two levels inside a product the homepage no longer sells. It is the answer to "why would anyone start using an RSS reader in 2026" — the single best onboarding surface in a category everyone assumes is dead — and the primary site barely points at it. The extra click was never the expensive part. The expensive part is that Feedly still owns the strongest argument for its own original product and has stopped making it.

What can't be read from outside is the arithmetic. Feedly may have measured reader traffic to three decimal places and concluded that promoting Discover buys nothing worth the homepage space. What's visible from here is narrower: the company decided its future is elsewhere, which is allowed, and the site still hasn't decided what to tell the people who are already here.

Take it further

The lens behind this teardown — can a reader tell what a thing does and whether it's still for them — is the UX Clarity framework, the same one we apply in a Full UX Audit. For how that scoring turns into prioritized fixes, read what a real UX audit looks like.

Sources: NN/g — Recognition vs Recall · Feedly News Reader · Feedly changelog (pages as captured 2026-08-07).

Has your product outgrown the audience your homepage still needs? Apply for a Full UX Audit →