The discipline, and where it stops
Website marketing is everything done to get the right visitors to a website and to get them to act once they arrive. That website marketing definition is deliberately narrow at both ends: it excludes work whose destination is not your site, and it excludes the product and pricing decisions that determine whether anyone wants what you sell.
Three things are routinely mistaken for it. Digital marketing is the wider field — a lead form completed inside LinkedIn, a podcast sponsorship, a call from a map listing. Search engine optimisation is one traffic source among several, and a site can rank well and convert nobody. Web development builds the thing; website marketing decides what it has to accomplish.
The useful boundary at the top is honesty about causation. Website marketing can make a good offer findable and easy to accept. It cannot make an uncompetitive offer competitive, and a company that has confused the two will spend years optimising pages while losing on price, product, or reputation.
Traffic, conversion, value — and which one to fix next
The arithmetic reduces to one identity: revenue from a site equals visitors, times the rate at which they take the action you count, times the value of that action once it works through the sales process. Three multiplied terms, not three added ones.
Because the terms multiply, the smallest is the constraint and improving anything else is wasted. Take the arithmetic on its own terms — this is illustration, not data. A site with a thousand relevant visits a month and a one-percent enquiry rate produces ten enquiries. Doubling traffic produces twenty. Doubling the enquiry rate also produces twenty, and on most sites it is cheaper, faster, and does not stop when the budget ends.
A fourth term ruins the first: qualification. Traffic drawn from audiences that cannot buy inflates visitors and depresses conversion at the same time, so the chart looks flat while cost per outcome climbs. Measure all three terms before funding any of them, then spend against the smallest.
Owned, earned, and paid, and the honest trade
Every visitor arrives through one of three mechanisms, and choosing between them is a choice about time, control, and what you keep.
| Type | What you are buying | When you stop | Main risk |
|---|---|---|---|
| Owned | Assets you control: indexed pages, an email list, documentation, your own events | It keeps working, decaying slowly | Slow to build; easy to under-resource for years |
| Earned | Other people's endorsement: coverage, references, citations, organic rankings | It persists but cannot be renewed on demand | Low control, unpredictable timing |
| Paid | Access to someone else's audience, priced per click, impression, or placement | It stops that day | Cost rises with competition; the audience is never yours |
The honest trade is that paid buys time and owned buys margin. A company needing pipeline this quarter should buy it; a company buying it every quarter for five years has rented its demand permanently. Cost structures differ too — paid is priced per unit of attention and inflates with competitor bidding, while owned assets carry a large production cost up front and almost none afterwards.
Why most plans are channel lists with no theory of the visitor
Open a typical website marketing plan and you find a table of channels with budgets and owners beside them. What is missing is any statement of who the visitor is, what question brought them, and which page answers it. That omission is why the plan produces activity and no result.
A plan with a theory of the visitor states five things before naming a channel.
- Who — the kind of company and the role, specifically enough to exclude people.
- What question they need answered, in their words rather than your product vocabulary.
- Which page answers it, and whether that page exists yet.
- What happens next — the one action the page is built to produce.
- How you will know, and the number at which you stop.
Channels then fall out of the answers instead of driving them. If nobody searches for the problem because they do not know it has a name, search will fail however much you spend, and the work is demand creation elsewhere.
Compounding assets versus rented attention
The most useful distinction in the discipline separates what accumulates from what you rent. Ask one question of any activity: if you stopped paying tomorrow, what would still arrive next month?
Compounding assets include pages answering durable questions that stay indexed for years, an email list you own outright, genuine references from other sites, and a brand recognised well enough that people search for your name. Rented attention includes every ad account, every platform's organic reach, and every audience assembled inside someone else's tool.
Rented mechanics also disappear without notice, and the record is not ambiguous. LinkedIn retired lookalike audiences on 29 February 2024, including the third-party API. Google stopped showing FAQ rich results on 7 May 2026 and dropped the reporting and testing support in the weeks after; HowTo rich results went earlier, with the documentation removed on 14 September 2023.
None of that argues against paid or platform work. It argues for knowing which part of your results you would keep.
Techniques that survived twenty years, and those that did not
This site has published on marketing corporate websites since 2003, long enough to watch many techniques arrive, get sold hard, and stop working.
Still works: publishing pages that answer questions buyers actually ask; linking those pages to each other sensibly; titles that describe the page; fast pages on real devices; an email list you built yourself; being useful enough that other sites reference you; and measuring against revenue rather than activity.
Stopped working, or never did: keyword density targets and the meta keywords tag; exact-match domains as a strategy in themselves; article directories, reciprocal link pages, and forum signature links; near-duplicate doorway pages for every city you serve; paid link networks; copy written for a crawler; and FAQPage markup added in the expectation of a rich result, which no longer produces one — though nothing documents it as harmful either.
Conversion testing sits in the middle: it works, but needs more traffic than most business sites have. Structured data remains useful for machine comprehension while Google states plainly that it is not required for its AI features.
A website marketing strategy you can actually operate
Most documents called a website and marketing strategy are budgets, calendars, or aspirations. An operable strategy is shorter and more uncomfortable, because it commits to some things and rules others out. It states six.
- The visitor you are trying to attract, and the ones you are content to lose.
- The set of questions you intend to own — a finite list, in the visitor's language.
- The page that answers each question, including those still to be written.
- The mechanism that brings people to it, chosen per question rather than per quarter.
- The action you count, and the business outcome it stands in for.
- What you have decided not to fund, written where whoever suggests it next quarter can see it.
Add two dates: a review cadence, and a kill date for anything that has not produced its stated outcome. A strategy that rules nothing out is a list. For company website marketing inside a larger organisation, add one clause: who decides.
Measurement, and when none of this is the priority
Report four things per traffic source: entrances, engaged sessions, key events, and the downstream revenue you can defend, over a window long enough to cover your sales cycle. Ignore bounce rate as a headline number, average time on page, and total pageviews.
Three facts change how the numbers read. GA4 derives sessions from events and session counts commonly differ from what Universal Analytics reported, so comparisons across that boundary are not like-for-like. GA4's retention setting caps user-level and key-event retention at fourteen months on standard properties, affecting explorations and funnel reports rather than standard aggregated reports — enough to matter when analysing a longer cycle. And GA4's behavioural modelling requires substantial sustained daily event and user volumes; most business sites do not qualify, so consent-related gaps stay gaps.
There are also cases where the discipline is the wrong investment. If your product sells exclusively through a distributor's catalogue, or your buyers genuinely do not use search to find suppliers, the site's job is credibility and due diligence, and money aimed at traffic acquisition would do more good spent on the relationships that generate the invitations.
Frequently Asked Questions
What is website marketing, in simple terms?
Website marketing is the work of getting suitable visitors to a website and getting them to take a useful action once there. It combines acquisition — search, paid advertising, email, referrals, and direct visits driven by other activity — with the work on the site itself that makes acting possible: pages, evidence, forms, speed.
It is narrower than digital marketing, which includes activity whose destination is a platform rather than your site, and broader than SEO, which is one acquisition channel.
Is it better to spend on more traffic or better conversion?
Measure both, then spend against whichever sits further from what is achievable. Because outcomes are traffic multiplied by conversion rate multiplied by value, the smaller term is the constraint and improving the other is largely wasted.
On established sites, conversion work is usually the better first bet: it is cheaper, it improves the return on every channel at once, and it does not stop when the budget does. The exception is a site with negligible relevant traffic, where the constraint is discovery.
How long does website marketing take to show results?
Paid channels produce data within days and revenue as fast as your sales cycle allows. Owned and earned work is slower: new pages must be crawled, indexed, and then accumulate enough query coverage to be judged, and the deals they influence may close a year later.
The reporting lag is set by your sales cycle, not by the channel. A company with a nine-month cycle cannot evaluate a content programme at ninety days.
What website marketing techniques are still worth doing?
The durable ones are unglamorous: publish pages that answer the questions your buyers ask, link them together properly, write titles that describe what is on the page, keep pages fast on real devices, build an email list you own, and earn references from sites your buyers already read.
What has not lasted: keyword density, article directories, reciprocal linking, doorway pages per city, paid link networks, and markup added purely to chase a search feature — FAQ rich results stopped appearing in Google Search on 7 May 2026.