There is a particular kind of business owner who spends six months obsessing over their product, another three months on their hiring process, and approximately one afternoon on their website. The site goes live, the domain is renewed automatically each year, and the whole thing is quietly forgotten until someone mentions it at a meeting.

This is not unusual. It is, in fact, the norm.

The problem is that the website doesn't know it's been forgotten. It keeps running. People keep arriving. And every day, it's either working for the business or against it — with no one watching to know which.

A website that takes four seconds to load loses roughly half of its visitors before they've read a word. A website that isn't mobile-optimised signals to Google that it doesn't belong at the top of search results. A website without clear navigation loses qualified visitors who simply give up.

The salesperson who never sleeps — and never converts

The analogy that makes this concrete: your website is your highest-volume salesperson. It meets more prospective customers than any human on your team. It works at 3am. It operates across every time zone simultaneously. It never takes a sick day, never misses a flight, never has an off month.

But unlike a salesperson you'd manage and coach, most websites are left entirely unsupervised. Nobody reviews the conversion rate. Nobody checks whether the contact form still works. Nobody asks why the homepage takes seven seconds to load on mobile, or why the pricing page hasn't been updated since a product change two years ago.

A mediocre salesperson on your team gets a performance review. Your website doesn't. And the cost of that neglect compounds quietly over time.

What "underperforming" actually looks like

The signs are rarely dramatic. Nobody calls to say your website cost them a sale. What you get instead is absence — the enquiry that didn't come, the comparison that was made in your competitor's favour, the recommendation that ended with "their website looks a bit old." These never show up in your CRM.

The clearer signals, if you know where to look, are in the data: a bounce rate above 70% on your service pages; average session duration under 45 seconds; near-zero traffic from organic search; contact form submissions that dropped 60% after a plugin update six months ago that nobody noticed had broken it.

Performance, in the technical sense, matters enormously here. Google's Core Web Vitals — a set of real-world performance metrics covering loading speed, interactivity, and visual stability — now directly influence your position in search results. A site that fails these metrics doesn't just feel slow; it ranks lower, which means fewer people ever find it to experience the slowness in the first place.

The conversation that changes things

The single most useful exercise a business owner can do before briefing any agency is to sit with their website and ask the question every first-time visitor would ask: does this make me want to get in touch?

Not "does this look nice?" Nice is subjective and mostly irrelevant. The question is functional: does this communicate clearly what you do, who you do it for, and why you're the right choice? Does it make the next step obvious and low-friction? Does it load in a reasonable time on a mobile phone with a mid-range data connection?

Most business websites, looked at honestly, fail this test. Not because anyone did a bad job, but because the brief was wrong from the start. Most website briefs focus on aesthetics — "we want something clean and modern" — rather than outcomes. The result is a website that looks like other websites in its industry, performs like other websites in its industry, and converts like other websites in its industry. Which is to say: not very well.

What a better brief looks like

A brief that generates a high-performing website starts with audience and goal, not with design references. Who arrives on this page, from where, and what do we want them to do? The answer to that question — specific, measurable, and honest — is the foundation everything else is built on.

From there: speed and performance requirements (not optional, not nice-to-have — specified and tested). Clear hierarchy of information. Calls to action that match the psychology of the buyer at each stage of the decision. SEO that reflects how your prospective customers actually search, not how you describe your own service internally.

None of this is especially complicated. What it requires is treating the website with the same seriousness you'd treat any other revenue-generating investment — because that is what it is. The businesses that understand this have websites that reliably generate enquiries and referrals. The businesses that don't have websites that are technically alive but professionally irrelevant.

The distinction, in the end, is not budget. It's intention.

When mobile apps first became widespread, a generation of commentators predicted the end of the domain name. Why type a URL when you could tap an icon? The prediction was wrong. Domains didn't fade — they became more valuable, because brand identity consolidated around them in ways that app stores never replicated.

Now the same question is being asked about AI. When ChatGPT or Perplexity can answer any question without routing the user to a website, what use is a domain?

The answer, this time, is more nuanced — and in some ways more urgent.

The citation economy

AI systems don't just answer questions. They attribute. When a language model recommends a service or cites a fact, it typically attaches a source — a domain name. The domain that gets cited is the one that benefits from the recommendation. The one that doesn't get cited might as well not exist for that query.

The domain isn't just the address anymore. It's the brand signal that AI systems use to identify and attribute authority to a source.

This creates a new dimension to domain strategy that barely existed three years ago. A domain name that is short, clean, semantically relevant, and clearly associated with a brand is more likely to be cited correctly by an AI system — because clarity reduces the chance of confusion, misattribution, or omission in a model's output.

The voice search dimension

Voice search compounds this. When a user asks a voice assistant for a recommendation and the assistant responds verbally, it reads a brand name aloud. A domain like mximzdigital.com becomes "mximz digital dot com" — which is fine, if it's pronounceable. A domain like mxmzd1g1t41.co becomes a noise that no one remembers and no one correctly types later.

The phonetic legibility of a domain — how clearly it communicates when heard rather than read — is a criterion that barely featured in domain acquisition advice until recently. Now, with voice interfaces embedded in phones, cars, speakers, and operating systems, it matters materially.

The .ai surge and what it means

.ai domain registrations grew 55% in 2025 alone. Part of this is branding — companies want to signal technological relevance. But part of it is genuinely strategic: Google now treats .ai domains as generic top-level domains, meaning they carry no SEO penalty relative to .com. For companies whose work is genuinely centred on artificial intelligence, the .ai extension provides semantic reinforcement that the .com equivalent would not.

This doesn't mean every business should rush to acquire a .ai domain. It means the choice of extension has become a strategic decision rather than a default — and one that deserves the same analysis you'd apply to any other significant brand asset.

What hasn't changed

The fundamentals that made domain strategy important in 2010 are still true: short is better than long, memorable is better than descriptive, unambiguous spelling is better than clever abbreviation. What has changed is the reasoning — the environment in which those principles operate has expanded to include AI systems, voice interfaces, and multi-platform attribution, all of which reward clarity in ways that traditional search alone never required.

The businesses that treated their domain as an afterthought in 2010 are the ones still stuck with awkward, hard-to-spell URLs that cost them visibility in local search and customer recall. The businesses that treat it as an afterthought in 2025 will face an equivalent penalty in the AI-mediated search landscape of 2030.

The question isn't whether domain names matter in the age of AI. It's whether you have a strategy for yours.

For the better part of two decades, the central preoccupation of anyone who wanted their business to be found online was SEO — Search Engine Optimisation. The goal was to rank as high as possible in Google's results for queries that prospective customers were likely to type. This was useful, measurable, and when done well, reliably rewarded.

SEO is not dead. But the environment it operates in has changed substantially — and a new discipline has emerged alongside it that many businesses haven't yet encountered: AEO, or Answer Engine Optimisation.

The difference between a search engine and an answer engine

A search engine returns a list of pages. The user selects one, visits it, reads it, and forms a view. The website gets the traffic. The brand gets the visibility. This is the world SEO was built for.

An answer engine — which is what ChatGPT, Perplexity, Google's AI Overview, and increasingly Bing Copilot are — returns a direct answer. The user reads it without visiting any website. The answer often cites sources, but the source gets mentioned, not visited. The traffic model is fundamentally different.

AEO is the discipline of making your content, your authority signals, and your brand presence legible to AI systems — so that when they synthesize answers, they draw on you and cite you.

What AI systems look for

Research into how large language models and AI search systems select content to cite reveals consistent patterns. Content over 2,900 words averages significantly more citations than content under 800 words. Content updated in the past three months receives nearly twice the citations of older, unchanged pages. Pages that use question-based headings and FAQ sections are disproportionately represented in AI-generated answers.

These patterns suggest that AEO rewards a different kind of writing than traditional SEO does. SEO has historically been associated with keyword density, backlink volume, and metadata optimisation. AEO rewards genuine depth, authoritative sourcing, clear structure, and — critically — the appearance of expertise that comes from content written by people who actually know their subject.

There is a painful irony in this for the brands that spent 2023 and 2024 flooding their websites with AI-generated content: those pages are now among the least likely to be cited by AI systems, precisely because they lack the specificity, depth, and original perspective that models are trained to identify as authoritative.

E-E-A-T and why it matters for AI citation

Google's concept of E-E-A-T — Experience, Expertise, Authoritativeness, and Trustworthiness — was originally a framework for human quality raters evaluating search results. It has become something more significant: a set of signals that AI systems use to determine whether a source is worth citing.

Experience means demonstrating that the author has first-hand knowledge of the subject. Expertise means demonstrating depth and nuance beyond surface-level information. Authoritativeness means being recognised by other sources — citations, mentions, links from credible domains. Trustworthiness means being consistent, accurate, and transparent about who is behind the content.

Building these signals is not quick work. It's also not technical work in the way that traditional SEO often was. It is fundamentally editorial — the slow accumulation of well-sourced, genuinely useful content written from a position of real knowledge.

The practical implications

For most businesses, the immediate AEO priority is simpler than it sounds: write clearly about what you actually know, structure your content so that AI systems can parse it easily (question-based headings help enormously), keep it updated, and ensure that your basic technical signals — structured data, schema markup, clear authorship — are in place.

The medium-term priority is authority building: getting your brand and your expertise mentioned in sources that AI systems trust — industry publications, professional directories, academic citations, news coverage. This is the new version of backlink strategy, translated into the citation economy that AI search runs on.

SEO is not going away. For a large proportion of queries, traditional search results still drive most of the traffic. But for the queries where AI systems now intervene — research questions, comparative queries, recommendations, how-to searches — AEO is the discipline that determines whether you exist in the answer or not.

The businesses that recognise this now, and begin building for it, will have a meaningful head start on the ones that wait until AI search is their primary traffic source to begin the conversation.

There is a category of business cost that accountants find difficult to capture and managers find easy to ignore: the cost of things that don't happen. The sale that wasn't made. The customer who arrived and left. The lead that found a competitor's website loading faster and never came back. These are real costs. They simply don't appear in any report.

Digital asset decay is one of the most common sources of these invisible costs. And it is almost entirely preventable.

What decay looks like in practice

A website built in 2022 using a popular CMS platform will, by 2025, have accumulated a significant backlog of plugin updates — each one a potential security vulnerability if left unpatched. The framework it was built on may have released several major versions, each introducing performance improvements that the unmaintained site is missing. The images that were appropriately sized for 2022 connection speeds may now be flagged as oversized in Google's Core Web Vitals assessment. The contact form plugin may have been deprecated by its developer and is now running on end-of-life code.

None of this announces itself. The website continues to look exactly as it did on launch day. Traffic numbers might not change dramatically for months. The decay is structural, not cosmetic — and by the time the visible signs appear (a security warning in Chrome, a sudden drop in search rankings, a contact form that silently fails to submit), the damage is already done.

The average time between a security vulnerability being discovered and it being exploited in the wild is 15 days. Most businesses check their website security far less frequently than once a fortnight.

The performance dimension

Google's Page Experience ranking signals — which include Core Web Vitals scores — are recalculated against actual user data. A site that was fast at launch may have accumulated enough additional scripts, unoptimised images, and heavy third-party embeds to have its performance score meaningfully degrade over 18 months. This affects search ranking directly. Lower ranking means less traffic. Less traffic means fewer enquiries. The causation is clear; the connection to the unmaintained website is not.

Mobile performance matters disproportionately here. The majority of web traffic now arrives via mobile devices — often on variable connection speeds. A website that loads in 2.1 seconds on a desktop in an office and 6.8 seconds on a mid-range phone in a city is, effectively, two different websites with very different conversion rates.

CRM drift

Customer Relationship Management systems require a different kind of maintenance, but the principle is identical. The workflows configured at implementation reflected the business processes of that moment. Businesses evolve. New products are added. Sales processes change. Staff turns over. Without periodic review, a CRM that was configured precisely for a business in 2022 is running on logic that no longer matches reality — automating the wrong things, missing the right ones, and gradually losing the trust of the sales team that is supposed to use it.

A CRM that the sales team doesn't trust is worse than no CRM. They route around it. Data stops being entered. Reporting becomes unreliable. Management decisions get made on bad information. The investment in the original implementation erodes without a single change being made to the system.

The case for a maintenance discipline

The argument against maintenance plans is almost always the same: "we don't have the budget." This argument implicitly assumes that the cost of maintenance is higher than the cost of not maintaining. In most cases, the opposite is true. A monthly security monitoring and update service costs a fraction of the emergency response to a successful attack on an unpatched site — and nothing compared to the reputational cost of customer data being compromised.

The better framing is that digital asset maintenance is not an expense. It is the amortisation of an investment. When a business spends money on a website, a CRM, or a mobile app, it has acquired an asset. That asset has a productive lifespan that is directly linked to how well it is maintained. Treating maintenance as optional is equivalent to buying a car and never changing the oil — the saving is real, and so is the eventual engine failure.

Most businesses that have experienced the emergency — the hacked site, the broken integration discovered mid-campaign, the app pulled from the store for OS incompatibility — subsequently maintain their digital assets with religious consistency. The smarter path is to reach that understanding without the crisis.

Most comparisons of Flutter and React Native are written by advocates of one or the other — which means they are written, in a structural sense, to reach a predetermined conclusion. This piece isn't that. The honest answer to which framework is right for your project is not written in advance; it depends on variables specific to your context. What follows is an attempt to lay those variables out clearly enough that the answer becomes obvious for your situation.

The shared premise

Both Flutter and React Native solve the same fundamental problem: building a mobile application that runs on iOS and Android without maintaining two separate codebases in two separate languages. This is a meaningful efficiency — native development in Swift for iOS and Kotlin for Android requires distinct teams or a team with fluency in both, and produces two applications that must be kept in sync as features evolve.

Cross-platform frameworks trade some degree of platform-specific optimisation for this efficiency. The question is how much each framework trades, and whether the trade matters for your use case.

Where Flutter has a genuine advantage

Flutter's core architectural decision — rendering everything through its own graphics engine, Skia (and now Impeller) — means that what you see on screen is not a wrapper around native components. It's Flutter drawing its own pixels. This has two consequences that matter in practice.

First, performance is highly consistent between iOS and Android. Because both platforms are using the same rendering engine, the same code produces the same visual output. There are no subtle differences in how a component renders on iPhone versus Pixel that need to be debugged and worked around.

Second, custom UI is dramatically more practical. If your design calls for non-standard components — unusual animations, custom chart types, bespoke interaction patterns — Flutter can implement them cleanly without fighting against native platform conventions. React Native, by contrast, often requires significant native module work for heavily custom UIs.

Flutter's rendering engine means you write once and get pixel-identical output on both platforms. For brands with strong visual identity requirements, this is not a minor consideration.

Where React Native has a genuine advantage

React Native uses JavaScript — specifically, a JavaScript runtime bridge to native components. This means that developers who already know JavaScript and React, the dominant stack for web development, can be productive in React Native relatively quickly. For organisations with existing web development teams who need to extend into mobile, this is a real practical advantage.

React Native's ecosystem is also older and, in some respects, broader. There are more third-party libraries, more Stack Overflow answers for edge cases, and more hiring candidates with direct experience. These are not trivial considerations for teams working in resource-constrained environments.

The Supabase and Riverpod dimension

When the question is not just "which framework" but "which framework with what backend and what state management architecture," the calculus shifts. Flutter with Supabase and Riverpod is a combination that has become increasingly coherent and powerful for a specific category of applications: data-driven, real-time, productivity-focused apps where reliability and testability matter.

Supabase — an open-source Firebase alternative built on PostgreSQL — provides authentication, database, storage, and real-time subscriptions in a single coherent backend. Its Flutter SDK is actively maintained and well-documented. Riverpod provides state management that is, compared to most of the alternatives, genuinely testable and free of the implicit global state problems that have plagued Flutter's earlier state management patterns.

This combination — Flutter, Supabase, Riverpod — is what you reach for when the app needs to be maintainable over a multi-year horizon, when the team working on it will change, and when you need the confidence that comes from a well-typed, well-tested codebase.

The honest answer

If your team knows JavaScript and React already, and you need to ship something quickly with a conventional UI: React Native is the pragmatic choice. If you're building something with a strong visual identity, complex custom components, or real-time data requirements, and you're investing in a team specifically for the project: Flutter is the stronger long-term choice. If you're building with Supabase as your backend and you care about testability and state management coherence: the Flutter, Supabase, Riverpod combination is among the best available stacks for production mobile development.

Neither framework is universally superior. The question is which set of trade-offs is right for your specific project, team, and timeline.

There is a particular kind of crisis that arrives without warning, accelerates fast, and is far more expensive to reverse than it would have been to prevent. In the context of business, online reputation incidents sit firmly in this category. A two-star review left unanswered for three months. A Reddit thread from a disgruntled former customer that climbed into the first page of branded search results. A Google Business listing with outdated information that has been generating complaints about opening hours that changed in 2023.

None of these start dramatically. All of them compound quietly. And the businesses that manage their online reputation well are, almost without exception, the ones that started before the problem — not after it.

What reputation actually means in 2025

Online reputation is no longer simply a collection of Google reviews. It is the aggregate of everything findable about a business across every platform where it has a presence — or where it has been discussed without its knowledge. This includes review platforms (Google, Justdial, Trustpilot, G2, industry-specific directories), social media mentions, news coverage, forum discussions, AI-generated responses to brand-name queries, and the search results that a prospective customer sees when they type your company name into Google before deciding whether to get in touch.

The breadth of this landscape is exactly why most businesses manage it poorly. There is no single dashboard. There is no single platform to monitor. And because the consequences of neglect are diffuse and delayed rather than immediate and concentrated, the urgency rarely materialises until something has already gone wrong.

A prospective customer who searches your brand name before making an enquiry is making a decision based on everything they find — not just your website. Most businesses have no idea what that experience looks like.

The review asymmetry problem

One of the most reliable patterns in online reputation is this: dissatisfied customers are significantly more likely to leave a review than satisfied ones. This is not because businesses serve more unhappy customers than happy ones. It is because unhappiness is motivating in a way that satisfaction rarely is. A customer who had a great experience thinks, vaguely, that they should leave a review. A customer who had a bad experience often feels compelled to.

The practical consequence is that a business with fifty happy clients and two unhappy ones may have a Google rating built almost entirely from those two. Without a deliberate strategy for generating reviews from satisfied customers — timing the request well, making it easy, asking the people most likely to respond positively — the default is a ratings profile that systematically misrepresents the actual quality of the service.

This is addressable. The businesses with strong, consistent ratings are not necessarily the ones doing better work than their competitors. They are the ones that have built review generation into their client journey as a deliberate, managed step.

Responding as a reputation signal

How a business responds to negative reviews tells prospective customers more about its character than the reviews themselves. A one-star review followed by a thoughtful, non-defensive response that acknowledges the issue and explains what was done about it frequently converts a reputational negative into a trust signal. The prospective customer doesn't just see the complaint — they see how the business handled it.

The opposite is equally true. A pattern of negative reviews met with defensiveness, deflection, or silence is read by anyone who finds it as a significant warning sign about how the business treats its customers when things go wrong.

Review response is not public relations work in the traditional sense. It doesn't require a communications team or a legal review process. It requires a consistent, genuine voice and a commitment to responding within a reasonable timeframe — which most businesses define as never, which is why it's an opportunity for the ones that do it well.

AI and the reputation landscape

The emergence of AI-powered search has added a new dimension to reputation management that most businesses haven't yet considered. When a user asks ChatGPT or Perplexity about a company — "what do people think of X?" or "is X a good agency?" — the AI system synthesizes an answer from whatever is findable about that company online. If what is findable is a thin website, two unanswered negative reviews, and a Google Business listing with incomplete information, that is the raw material from which the AI answer is built.

Businesses with strong, well-maintained online presences — coherent information across platforms, responded-to reviews, published case studies, consistent brand voice across social media — give AI systems better raw material to work with. The result is more accurate, and typically more favourable, AI-generated representations of their brand. This is not gaming the system. It is ensuring that the system has accurate information to work with.

The audit that most businesses have never done

The single most useful first step in reputation management is one that costs nothing beyond time: search your own brand name from an incognito browser, in the city where your customers are, and look honestly at what comes up on the first page. What is the review rating? Is the Google Business information correct? Is there anything in the results that you wouldn't want a prospective customer to see? What does the Knowledge Panel say, if there is one?

Most business owners, doing this exercise for the first time, discover things they didn't know were there. This is the starting point — not a crisis, but a baseline. Reputation management from that baseline is methodical rather than reactive: claiming and completing all relevant business listings, establishing a review generation process, monitoring for new mentions, and building the kind of consistent online presence that reflects the business you actually run.

The businesses that do this well tend to share one characteristic: they started before they had a problem. The ones that learn the hard way wish, without exception, that they had.