The question comes up in almost every project conversation: does our website really need a blog? The honest answer is that it does only if the blog answers the questions your customers are already asking. A blog documenting company outings and anniversaries costs time and returns very little. A blog answering the twelve questions that surface in every sales conversation, on the other hand, works for you in the evening, at the weekend and long after the day it was published. This article looks at when a company blog pays off, how much work a single post really takes, which publishing rhythm is realistic for a business without an editorial team, and why maintaining what you already have often beats writing something new. It also places AI where it belongs: helpful for drafts and structure, but not a substitute for expert review.
When a blog pays off for a small business
A blog does not pay off through the number of posts but through the questions it answers before anyone picks up the phone. 99.3 percent (Federal Statistical Office of Germany) of all companies in Germany are small and medium-sized businesses, and across the European Union the figure is around 99 percent (European Commission) — businesses, in other words, where nobody writes text for a living. At the same time, roughly 89 percent (Federal Statistical Office of Germany) of companies with internet access run their own website. The difference between a website that merely exists and one that produces enquiries rarely lies in the design. It lies in whether visitors find what they need before making a decision: how does this work, what does it cost roughly, which documents are required, how long does it take? A service page can answer questions like these only up to a point without becoming overloaded. The blog is the place for depth. It pays off when it does three things: it answers questions that arise before the first contact, it takes work out of the initial consultation because prospects arrive prepared, and it gives your site reasons to appear for specific wordings that would not fit on a service page. To see how this connects with the rest of the build, take a look at our services around building and running a website.
A blog does not replace a solid service page
The honest effort behind one post
The workload is rarely discussed openly, and that is precisely where many company blogs fail. A usable specialist article of 800 to 1200 words does not appear in twenty spare minutes, but it does not swallow two working days either. Three to five hours is realistic, provided the topic, the outline and the images are ready and a draft serves as the starting point. Most of that time goes not into writing but into two other steps: choosing the topic and reviewing the facts. The choice of topic decides whether the text reaches anyone at all. The review decides whether the statements hold up in a customer conversation. Skipping those two steps saves an hour and risks a post that creates false expectations and generates follow-up questions later. It helps to record the effort honestly once: anyone who notes how long the first post actually took will plan the second far more realistically. In practice the time needed drops noticeably from the third or fourth text onwards, because structure, tone and filing routines settle in. For the technical side — from input through preview to the statically delivered page — see the overview of how the process works behind the interface.
| Work step | Typical time needed | Who handles it |
|---|---|---|
| Choosing and narrowing the topic | 15 to 30 minutes | The business, based on real customer questions |
| Outline and key statements | 20 to 40 minutes | The business, or an AI draft as a proposal |
| Writing the first version | 60 to 180 minutes | Draft generated, revision inside the business |
| Expert review and correction | 30 to 45 minutes | A qualified person, not delegable |
| Selecting and describing images | 15 to 30 minutes | The business, alt text suggested |
| Title, description, internal links | 15 to 20 minutes | Editing interface with suggestions |
| Publishing and visual check | 10 minutes | The business, with an eye on mobile |
Topics come from real customer questions
The most workable topics are already sitting in your inbox. Around 92 percent (Federal Statistical Office of Germany) of people aged 16 to 74 in Germany use the internet, and a considerable share of that use is research before a decision. People rarely search using industry vocabulary; they search in everyday language, asking about cost, deadlines, prerequisites and comparisons. Those exact phrasings reach you every day free of charge — through the contact form, on the phone, in consultations. Collect them unsorted in a single list for four weeks without judging them. Then rank them by frequency: anything asked three times or more becomes a post, everything else stays a note. This approach has two advantages. First, you write about things you master professionally, which shortens the review time. Second, you match your customers' vocabulary instead of using terminology nobody types into a search field. If your enquiry form is already being evaluated, it delivers the topic list almost by itself — the post on turning visits into enquiries with contact forms describes which fields make that possible. A useful side effect: you quickly spot which questions belong on the service page and need no separate post at all.
- Review the last fifty form enquiries and write down recurring questions word for word
- Keep a log of phone calls for a week: which question comes up before almost every appointment?
- Analyse quotes and follow-ups: where do misunderstandings about process, duration or price arise?
- Ask the team: which explanation do colleagues repeat in every second conversation?
- Collect seasonal triggers: what gets asked in spring, in autumn, before the turn of the year?
- Evaluate internal search terms on your own site, if a site search exists
- Assign every question to one of three categories: decision, process or cost
Decision questions
Which option fits my situation? These posts compare alternatives factually, name prerequisites and limits, and hand over to the matching service page at the end rather than selling in every paragraph.
Process questions
How does this actually work? Texts that describe the process in steps win here: who does what, which documents are needed, and how long each stage typically takes from start to handover.
Cost questions
Roughly what does it cost? Publishing price ranges together with the factors that move them, instead of avoiding the subject, filters out unsuitable enquiries and saves both sides time.
Rhythm: dependable beats plentiful
The typical arc of a company blog looks like this: six posts in four weeks, then eight months of silence, followed by a guilty conscience. That arc has little to do with a lack of discipline and everything to do with planning that ignores daily reality. In a business with a full order book, one post per month is a realistic target, and in peak season one per quarter may be all there is. What matters is not the volume but that the rhythm fits the business and is actually kept. Twelve reviewed posts a year add up to an archive of around forty texts after three years, and that carries noticeable weight — provided it is maintained. What works in practice is fixing topics in advance and reserving time slots rather than waiting for an opportunity: half a morning per month, blocked in the calendar like a customer appointment. Pulling seasonal topics forward buys breathing room too: the post about the heating season is written in August, not during the first cold snap. For businesses with a regional catchment area it is also worth reading how the pieces fit together in the article on local visibility for trades and hospitality: a specialist post with a local reference works differently from a generic guide with no location at all.
| Model | Posts per year | Effort per month | Suited to |
|---|---|---|---|
| One post per month | 12 | 3 to 5 hours | Businesses without an editorial team, long-term build |
| Every two weeks | 24 | 6 to 10 hours | A named owner and a well-filled topic list |
| Quarterly focus | 4 to 8 | 2 to 3 hours | Seasonal trade and project-based work |
| Whenever there is time | unpredictable | hard to plan | Rarely sustainable, the archive goes stale |
One post a month that somebody is professionally accountable for tends to work longer than ten texts nobody has read back.
Quality over quantity: what makes a post hold up
The craft standard for a specialist post is modest, yet it is frequently missed. Around 71 percent (HTTP Archive Web Almanac) of the pages examined serve a meta description at all; for the rest, chance decides which snippet of text appears in search results. On accessibility the picture is starker: roughly 95 percent (WebAIM Million) of the home pages examined show automatically detectable failures against common success criteria of the Web Content Accessibility Guidelines (W3C). For blog posts these are usually the same three issues: missing alt text, skipped heading levels and links without a meaningful anchor. Fixing them costs a few minutes per post and decides whether a text stays usable with a screen reader or in poor lighting. On content, one simple rule applies: one question per post, visible in the title, answered in the opening paragraph, evidenced in the rest. Anything else that seems interesting belongs in its own text. For images it is worth reading about formats and alt text for website images — in blog posts, images are often the heaviest payload and at the same time the least well described element on the page.
- One question per post, recognisable in the title and answered in the opening paragraph
- Concrete details from your own operation instead of generic phrasing
- Figures with a source and a date so they stay verifiable and easy to update later
- A meta description that names the benefit instead of repeating the title
- Subheadings in a sensible order, without skipping levels
- Images with meaningful alt text and fixed dimensions in the markup
- One clear next step at the end: the matching service page or a way to get in touch
Volume without review damages the archive
Maintain the archive instead of writing something new
The most frequently overlooked lever in a company blog is not the next post but the one before last. A text that already gets found can be improved substantially with one or two hours of revision: updated price ranges, new legal conditions, added practical examples, sharper subheadings. Producing something new costs three to four times as much and starts from zero. That is why every editorial plan needs a second column for revisions next to the column for new topics. Twice a year is usually enough: sort the archive by age and traffic, pick out the posts with outdated details and work through them. It matters that the modification date is shown visibly and carried in the structured data, so the revised state is recognisable outside the page as well. Where two posts cover the same ground, merge them and redirect the old path to the new one. The article on running a website relaunch without losing rankings describes how to plan such redirects cleanly. The same discipline applies in miniature to every merged blog post.
- Twice a year, list all posts and sort them by age, traffic and how current they are
- Revise the posts with outdated details first, not the ones with the fewest visits
- Check prices, deadlines, named contacts and legal notes against the current position
- Add new findings and remove superseded sections instead of leaving them in place
- Show the modification date visibly and carry it in the structured data
- Merge posts that overlap in substance and redirect the old path permanently
{
"@context": "https://schema.org",
"@type": "BlogPosting",
"headline": "Retrofitting underfloor heating: process, timeline and cost range",
"datePublished": "2026-03-04",
"dateModified": "2026-07-15",
"inLanguage": "en-GB",
"author": { "@type": "Organization", "name": "Sample Trades Ltd" },
"publisher": { "@type": "Organization", "name": "Sample Trades Ltd" },
"mainEntityOfPage": "https://www.sample-trades.example/blog/retrofit-underfloor-heating/",
"about": "Renovation of existing buildings",
"isPartOf": { "@type": "Blog", "name": "Notes from the workshop" }
}Updating often beats producing something new
Connecting posts with service pages and enquiries
A post without a follow-on is a dead end. Answering a question and then leaving the reader alone wastes the moment when interest is at its peak. Every post therefore needs at least one link to the service page it belongs to, and each service page in turn needs links to the two or three posts that go deeper on typical follow-up questions. This mutual linking has two effects: it moves people forward, and it makes the thematic relationship obvious to search engines. Machine-readable markup is worth adding as well: only about 41 percent (HTTP Archive Web Almanac) of pages requested on mobile contain structured data at all, even though posts benefit from it directly through author, publication date and modification date. The article on structured data and standing out in search results explains which markup suits which page type. At the end of each text there should then be a single clear next step instead of four competing calls to action. Which service page fits usually follows from the question itself — an overview of the possible destinations is given in the detailed description of our services.
From post to service
Each text closes with a link to the service page that belongs to the question it answered. One link per post is enough; several offers side by side weaken the decision rather than supporting it.
From service to post
The service page links to two or three posts that go deeper on common follow-up questions. That keeps the page itself lean while still covering everything a prospect wants settled before enquiring.
From post to enquiry
The way to get in touch belongs visibly in the text, not only in the footer. A short prompt that refers to the topic tends to work better than a generic invitation to contact the company.
Placing AI honestly: drafting yes, sign-off no
Around 20 percent (Federal Statistical Office of Germany) of companies in Germany use AI technologies, and among smaller businesses the share is lower. For editorial work this mainly shifts where the effort sits: the first draft is no longer the bottleneck, and the review becomes the actual contribution. A language model can propose a sensible outline, pick up the tone of existing pages and deliver a complete draft that works as a basis. What it cannot do is know which deadlines apply in your region, which price range you actually quote, or which exception you made on the last project. That is exactly where the typical errors appear — plausible-sounding details with no basis. Expert review therefore stays with the business, without exception. XICflow produces post drafts in the voice of your own website, including outline, subheadings and suggested alt text; revision and sign-off then happen with you. For the wider picture of how this collaboration works, see the article on creating a website with AI in a small business. If you would rather look at finished results first, there are examples among the demo websites.
| Work step | What an AI draft delivers | What the business checks |
|---|---|---|
| Outline | A structure of question, process and next step | Whether the order matches your own practice |
| First version | Phrasing in the style of existing pages | Terminology, tone and how binding the statements are |
| Figures and deadlines | Placeholders flagged as needing evidence | Every detail against a source, price list or contract |
| Legal notes | General wording without reference to a case | Review by qualified people before sign-off |
| Images and alt text | Suggestions based on the detected image content | Whether the description matches what is actually shown |
| Internal links | Suggestions for thematically related pages | Whether the destination matches the intent of the post |
A review loop that has proved itself
Measuring what actually counts
The most common measurement mistake in a company blog is staring at page views. For a business with a regional catchment area, a post with 120 views a month that produces three qualified enquiries is worth more than one with 4000 views and no contact at all. What helps is a short list of metrics that traces the path to the enquiry, plus a fixed review date — quarterly is enough, monthly only produces noise at small volumes. Patience matters too: a new post typically needs three to six months before its position settles. Deciding after four weeks whether the blog is worthwhile means deciding too early. One decidedly unstatistical question in the first conversation helps as well: where did you read that already? When prospects refer to a post unprompted, that is a more reliable signal than any curve. Record the actual time spent per post as well; it is the basis for planning next year's rhythm realistically. To see how editorial work maps onto the feature set, look at the overview of what is included and what it costs.
- Enquiries where a post was viewed along the way, rather than raw page views
- Search queries a post is found through, and how they develop across quarters
- Time on page and scroll depth as an indication of whether the question was really answered
- Clicks from the post to the linked service page as a measurable handover point
- How often prospects mention a post themselves in the first conversation
- The share of posts that have been revised within the past twelve months
- Actual time spent per post as the basis for next year's planning