Marketing for website owners tends to start from the wrong end. Somebody builds a site, notices it is quiet, and starts buying activity: posts, ads, backlinks, a newsletter, a redesign. Activity is easy to sell because it is easy to describe and easy to invoice. What actually changes the number of enquiries is a much shorter list, done in a particular order, and most of it is unexciting. This page sets out that order, explains which parts an agency should be doing and which you should keep in house, and gives you the questions that separate a proposal built around your business from a package built to be sold repeatedly.
First, make the site findable and fast
Before any campaign, three technical questions decide whether anything else can work. Can search engines reach and index your pages, does each page make plain what it is about, and does the site load quickly on a mid range phone on a mobile network. Google's own starter guide is direct that being crawlable and having clear, descriptive titles and headings are foundational, and that changes take time to show. On speed, the Core Web Vitals metrics documented on web.dev give you a concrete set of measures for loading, interactivity and layout stability, which turns a vague complaint about the site feeling slow into a specific target an agency can be held to. This part is finite work. It has an end. If a proposal offers technical optimisation as a permanent monthly line item without saying what will be fixed and when it will be done, you are being sold a subscription to a project.
Second, build pages for what people actually search
The single biggest cause of a quiet site is that its pages are written about the company rather than about what buyers type. A services page listing everything you do competes for nothing in particular. A page about one specific job, written for someone who has already decided they need that job done, competes for a query with real intent behind it. That is why the useful first exercise is not a content calendar, it is a list of the things customers ask for by name, sorted by how likely each is to end in a sale. Then one page per item, written by somebody who understands the work, with the evidence a buyer needs on the page: what it involves, what it typically costs or what drives the cost, who it suits, and what happens next. Google's guidance on helpful, reliable, people first content asks whether a page demonstrates first hand expertise and whether a reader leaves feeling they need to search again. Most thin service pages fail on the second question, and they fail whether a human or a model wrote them.
Third, remove the friction between reading and enquiring
A surprising share of what looks like a traffic problem is a conversion problem. Phone numbers that are not tappable on mobile. Forms with eleven fields when three would do. No indication of price, service area, or whether the company even takes work like yours. Contact details buried on a separate page. Enquiries landing in an inbox nobody reads until Thursday. None of this is glamorous and none of it needs an agency, but it routinely doubles the useful output of the traffic you already have. Test it yourself: open your own site on your phone, pretend you are a customer, and try to hire you. Time it. Then do the same on the two competitors who outrank you and note what they tell you that you do not. Buyers evaluating combined digital marketing and SEO services should ask specifically who fixes this layer, because it falls between the design contract and the marketing retainer often enough to sit undone for a year.
Judging a proposal, and judging the results
A proposal worth accepting names the outcome, the pages or campaigns that will be produced, the cadence, who does the work, and what you will be able to see each month. A proposal to avoid lists channels and hours without connecting either to a decision your customer makes. Ask three questions of any candidate. What will exist at the end of month three that does not exist today. Which of these tasks would you drop first if the budget halved, and why. And who owns the analytics, Search Console and ad accounts (the answer should be you, with them granted access). On measurement, agree the definition of a result before the first invoice: a form fill, a booked call and a signed job are three different things, and a great many disagreements at month six trace back to nobody writing down which one was being counted.
Questions people ask about marketing for website
Should I fix the website or start advertising?
Fix the obvious conversion problems first, because advertising multiplies whatever the site already does with its visitors, including doing nothing with them. Beyond that, ads and search work answer different needs: ads buy visibility this week and stop when the budget stops, search work compounds slowly and keeps working. Most small businesses end up needing both, in that order.
How much content does a small site actually need?
Fewer pages than most content plans propose, each one substantially better. A page for every distinct thing you sell, written properly, plus a small number of pages answering the questions buyers ask before they commit. A stream of short posts on topics everyone has already covered adds pages without adding evidence that you know the work.
Can I do this myself?
The conversion fixes and the writing about your own trade, yes, and you will do the second better than most agencies because you know the work. Technical repairs, structured data and campaign management are where outside help pays for itself. Be honest about the constraint that actually binds, which for owner operators is almost always time rather than knowledge.
How long before a site starts producing enquiries?
Paid traffic can produce them the same week. Organic work usually takes months and its first visible signs are impressions and rankings on target queries rather than enquiries. Watch leading indicators from month one, pages shipped and indexation among them, so you find out early if the problem is that very little is actually being published.