What we are learning right now is changing how we build websites, write content, and advise clients, and we think that is worth saying out loud. Agencies have a bad habit of presenting themselves as finished. Nobody in this industry is finished. The tools shift, search behavior shifts, and the honest position is that the work only stays good if the learning never stops.
This post is a snapshot of where our attention sits today: what has changed, what we got wrong, what we are testing, and how those lessons show up in the projects we deliver.
Key Takeaways
- Search behavior is shifting toward answers rather than links, which raises the value of brand, clarity, and original expertise.
- Performance and accessibility are design decisions, not developer cleanup at the end of a build.
- Fewer, stronger pages consistently outperform large sites full of thin ones.
- The fastest learning comes from shipping, measuring, and revising real client work, not from collecting theory.
- We are unlearning as much as we are learning, and letting go of old habits is the harder half.
Why We Talk About Learning Out Loud
It would be easier to publish case studies and stay quiet about the rest. We do not, for two reasons. First, clients deserve to know how their partner thinks, not only what their partner sells. Second, writing down what changed forces us to be specific. Vague learning is just a feeling. Written learning becomes a checklist somebody can hold us to.
If you are evaluating agencies, this is a fair question to ask any of them. What have you changed your mind about in the last year? A team with no answer either is not paying attention or is not being straight with you.
What We Are Learning About How People Search
The top of the funnel looks different than it did even two years ago. More questions get answered on the results page or inside an AI assistant, which means fewer informational clicks reach the site that did the work. That is not a reason to stop publishing. It is a reason to publish differently.
Answers Get Summarized, Expertise Gets Cited
Generic explainer content was always replaceable. Now it gets replaced automatically. What still earns a click is the thing a summary cannot flatten: a real process, a firm opinion, a specific number from real work, a comparison somebody actually ran. We read the Google Search Central blog directly instead of waiting for the recap, because the recaps lose the nuance that matters.
Branded Search Is the Metric We Watch Now
When people search your company name, they have already decided you are worth checking out. That number says more about marketing health than raw session counts do. We track it for clients now, and we treat a rising branded search line as evidence that the rest of the work is landing.
Local Still Behaves Differently
For service businesses, the map results and the business profile often matter more than where the website ranks. Reviews, accurate categories, and real photos move the needle. This has been true for a while, and we keep meeting owners pouring effort into blog posts while their profile sits half finished.
What We Are Learning About Web Design
Design conversations used to end at how it looks. Increasingly they start at how it performs and who it excludes.
Speed Is a Design Decision
Most slow sites were made slow during design, not during development. Five typefaces, an autoplaying video header, a slider nobody clicks, and a stack of third-party scripts add up before a single line of code gets written. We now discuss page weight during the design phase, and we say no to more things than we used to.
Mobile Is Not a Breakpoint, It Is the Default
The majority of first visits happen on a phone for nearly every client we work with. Designing desktop first and squeezing it down still produces cramped navigation, tiny tap targets, and forms nobody finishes. Our approach to mobile-first web design exists because the reverse order kept failing in measurable ways.
Accessibility Improves Everything Else
Building for accessibility used to feel like a constraint. In practice it makes sites better for everyone: stronger contrast, clearer labels, sensible heading order, keyboard navigation that actually works. Much of it is straightforward once it becomes a habit instead of a retrofit, and the guidelines are public rather than mysterious.
Fewer Pages, Better Pages
We keep finding sites with sixty pages where twelve would perform better. Thin pages dilute authority, confuse navigation, and create maintenance nobody has time for. Consolidating is often the highest-value thing we do in a redesign, and it is almost never what the client expected to hear.
What We Are Learning About Content
The volume era is over for most businesses. Publishing four mediocre posts a month is worse than publishing one that somebody bookmarks.
- Depth on a narrow topic beats coverage of a broad one, especially for smaller sites.
- Updating an existing post usually outperforms writing a new one on the same subject.
- Original detail is the moat: your process, your tradeoffs, your reasons for saying no.
- Formatting is content. If nobody can scan it, the quality of the writing stops mattering.
- Writing for one specific person beats writing for an imagined average reader every time.
Research on reading behavior has been consistent for a long time, and the Nielsen Norman Group article library is still where we send people who believe users read every word. They do not. They never did.
What We Are Learning About Social
Social changed shape again, and pretending otherwise wastes client money. A few things we have adjusted.
- Reach follows relevance more than follower count, which is good news for smaller accounts.
- Consistency in one format beats scattered experiments across five formats.
- Short video keeps winning attention, and the production bar is lower than most owners fear.
- Comments and replies do more for a small local business than another polished post does.
- Social rarely closes a sale. It shortens the distance between a stranger and a phone call.
We also stopped promising that social would replace other channels. It supports them. That is a smaller claim and a far more honest one.
How We Actually Learn
Learning that never touches real work is entertainment. Our process is unglamorous on purpose.
Ship, Measure, Revise
Every project gets a review after launch. Not a celebration, a review. What did the analytics do, where did users get stuck, which assumption turned out wrong, and what would we do differently next time? Those answers go into a shared document that shapes the next build.
Read the Source, Not the Recap
Documentation, official blogs, and primary research beat secondhand hot takes. Most of the panic that circulates online traces back to somebody misreading one sentence. Reading the original costs ten extra minutes and prevents a quarter of wasted effort.
Argue Internally, Agree Externally
We push back on each other during the work. Somebody has to defend the decision that adds a script, or the heading that says nothing, or the page nobody can justify. Disagreement early is much cheaper than a redesign later.
What We Are Unlearning
The harder half of learning is dropping the things that used to work.
- That more pages automatically means more traffic.
- That a homepage carousel is a reasonable way to handle competing priorities.
- That word count targets tell you anything useful about quality.
- That a redesign fixes a business problem the old site did not cause.
- That clients want options. Most want a recommendation and the reasoning behind it.
What We Are Testing Right Now
A few things are still open questions for us, and we would rather say so than perform certainty.
- How much traditional blog traffic still converts once AI summaries handle the simple questions.
- Whether structured, question-led page formats earn more citations than long narrative guides.
- How small teams keep short video consistent without it consuming the entire week.
- Where the real payoff sits between a full redesign and a series of targeted page rebuilds.
We will report back when we have something firmer than an opinion.
Why This Matters for Your Business
None of this is academic. It shows up in the invoices and in the results.
- You get fewer pages built and more attention paid to the ones that convert.
- You get performance and accessibility handled during design instead of billed later as a fix.
- You get advice about where not to spend, which is the part most agencies skip.
- You get the reasoning explained, so your team can carry the work forward without us.
You can read more about how we work and what we care about on our about page, though honestly this post probably says more about it than that one does.
Frequently Asked Questions
How often does your approach actually change?
The fundamentals rarely move. Clarity, speed, useful content, and a clear next step have worked for years and still do. What changes is execution: which formats earn attention, how search surfaces answers, what platforms reward. We reassess execution constantly and treat the fundamentals as settled until something genuinely proves otherwise.
Does chasing new trends put my project at risk?
It would, which is exactly why we separate the two. New techniques get tested on internal projects and low-risk pages before they touch a client build. If something is not stable, well documented, and supported broadly enough for your audience, it does not ship on your site. Curiosity belongs in the lab, not in production.
How do you decide what is worth learning?
We ask whether it changes a decision. A tool that shaves ten minutes off a task is nice. A shift in how customers find businesses is a priority. Anything that only affects how the work feels internally goes to the bottom of the list, no matter how much noise it is making online.
Can I expect my agency to explain all of this to me?
Yes, and you should insist on it. A good partner can explain a recommendation in plain language, including what it costs, what it risks, and what happens if you skip it. If the explanation only works when you already know the jargon, that is a communication problem, not a complexity problem.
Where We Go From Here
The honest summary is that we are paying attention to fewer things more carefully. Fewer pages, fewer channels, fewer tools, more depth in each one. Search is rewarding real expertise, design is being judged on performance and access, and content is being judged on whether anyone finishes reading it. All of that points in the same direction.
If any of this lines up with what you are wrestling with, we are glad to compare notes. Start a conversation with our team and tell us what is not working. We will tell you what we would try and why.
