Skip to content
PageSpeed 100 as the delivery default
Website-Grundlagen

New Website: When the Investment Really Pays Off

Ten triggers in a self-test, a scoring scale from maintain to rebuild, and the maths behind it: how many extra orders a new website has to carry a year.

15 min read EntscheidungRelaunchChecklisteHandwerk

The question rarely comes out of nowhere. There is usually a trigger: a customer mentions in passing that she could not read the price list on her phone. An application fails to arrive although the vacancy has been open for weeks. Someone calls and asks about a service the company stopped offering two years ago. From that moment the question is on the table: is it enough to patch up the existing website, or is a rebuild the more honest route? This article answers that question deliberately without technical detail. It provides a self-test with ten triggers, a scoring scale with three bands, a sober list of what a website can and cannot deliver, and finally the maths behind it: how many additional orders per year an investment has to carry in order to pay for itself. Anyone still unsure afterwards at least has a number instead of a gut feeling.

New website: is the step worth it?Ten triggers, a scoring scale and the maths behind it94 percent run their own siteTen triggers in the self-test6 of 10 applyNo enquiry through the site for monthsBarely usable on a phoneContent or prices out of dateNobody in-house can change anythingLegal pages missing or outdatedPrevious provider unreachableLong load times on mobile networksNo access to domain or hostingServices have changedCompetitors visible, your firm notScoring scale6 of 10triggers apply0 to 2: keep maintaining3 to 5: revise areasfrom 6: rebuild is cheaperThe maths: what it must carryOrder valueContributionOrders/year400 euros120 euros151,200 euros360 euros54,000 euros1,200 euros2Example: 1,800 euros a year, 30 percent contribution marginThe self-test does not replace a case-by-case reviewSources: Bitkom, Destatis, Think with Google

Between patching up and rebuilding

The question of principle has long been settled in most companies. 69 percent (Federal Statistical Office of Germany) of enterprises in Germany had their own website in the most recent survey; among those with 250 or more employees it was 99 percent (Federal Statistical Office of Germany), among those with one to nine people around 65 percent (Federal Statistical Office of Germany). In the skilled trades the share is even higher: 94 percent (Bitkom) of the businesses surveyed name their own website as a measure for being visible online, followed by directory entries at 88 percent (Bitkom). In Bavaria, 92 percent (Bavarian State Office for Statistics) of enterprises with at least ten active persons had a website.

So the practical question is almost never „website yes or no", but: does the existing one still carry its weight? And that question is rarely asked cleanly. Instead, either nothing happens at all because the effort is unclear, or some small thing gets repaired every few months without anything changing in the outcome. Both cost money, just less visibly than an invoice does.

A lack of capacity is a key reason for the hesitant digitalisation of German skilled trades: by their own account, close to three quarters of trade businesses simply have too much on their plate to engage more deeply with digitalisation.

Bitkom, study report on the digitalisation of the skilled trades (survey of 504 businesses, 2025)

The reticence has understandable reasons. 89 percent (Bitkom) of trade businesses view digitalisation positively in principle and only 6 percent (Bitkom) see it as a risk — yet 69 percent (Bitkom) name high investment costs as a central barrier, 72 percent (Bitkom) say they are too busy with day-to-day work, and 59 percent (Bitkom) consider digital applications economically worthwhile only for larger firms. Anyone who has to decide under these conditions does not need a debate about technology, but a handful of verifiable indicators.

The decision is made before the quote

The most expensive mistake in website projects does not happen during the build but before it: a quote is requested without anyone being clear about which problem is supposed to be solved. Then price decides, not need. The self-test below reverses the order — first the findings, then the number, then the conversation.

The self-test: ten triggers

The following ten points are not matters of taste. Each of them can be answered with yes or no without knowing anything about technology. The ground rules: what counts is the state of the past twelve months, not the intention. Each applicable point counts as one tally. If you are unsure, do not count it — the test should neither flatter nor dramatise.

  • For months, not a single enquiry has demonstrably come in through the website.
  • On a phone the site can only be used by zooming, swiping and aiming carefully.
  • Content, prices, contact persons or opening hours are no longer up to date.
  • Nobody in the company can change a text, a price or a photo themselves.
  • Imprint and privacy policy are missing, incomplete or years out of date.
  • The previous service provider can no longer be reached or no longer replies.
  • The page takes noticeably long before anything readable appears.
  • There is no own access to the domain, the hosting or the editing system.
  • The range of services has changed, the website still shows the old state.
  • For the usual searches competitors appear, your own company does not.

Six of these ten points weigh more heavily in practice than the remaining four, because diligence cannot compensate for them. They describe situations in which every maintenance task becomes more expensive than the matter is worth.

No enquiries through the site

If nothing has arrived through the form, the phone link or the email address for months, there are two explanations: nobody arrives, or nobody gets through. Both can be checked before money is spent. The article on enquiries from the contact form covers the second possibility in detail.

Not usable on a phone

When buttons are too small, tables run off the screen or phone numbers cannot be dialled, the site loses precisely where most visits happen. How to check this without any tools is shown in the article on mobile usability.

Visibly outdated content

A price from the day before yesterday, a team that no longer exists in that form, a note about last year's campaign: such spots devalue the rest of the site as well, because visitors conclude that everything else is old too.

Nobody can change anything

When every change is an order placed with somebody else, the small corrections simply do not happen — and those are exactly what keeps a website alive. What a workable routine looks like is described in the article on ongoing website maintenance.

Legal pages missing or outdated

Imprint and privacy policy are not optional extras. If they are missing or no longer match how the site actually operates, that is a risk which patching rarely resolves cleanly. The article on imprint and privacy policy sorts out the duties.

No access to domain or hosting

Without your own access, the address effectively belongs to somebody else. This is the one point on the list that has to be resolved immediately regardless of the score — background in the article on domain and business email.

The four remaining triggers are less dramatic but measurable. For loading times there is a well-known order of magnitude: if a mobile page takes longer than three seconds to load, an average of 53 percent (Think with Google) of visits are abandoned. And the state of the web as a whole shows this is not a fringe problem: only 48 percent (HTTP Archive Web Almanac) of pages measured on mobile pass all three metrics for loading, responsiveness and layout stability.

Long loading times

The test takes a minute: open the page on your own phone over the mobile network, not on Wi-Fi, and count silently to three. If nothing readable is visible by then, the point counts. How to solve this structurally is described in the article on static delivery.

Provider no longer reachable

When nobody answers any more, every change turns into reconstruction work: where does the site live, who owns the address, who can renew the certificate? The point counts as soon as the last reply is more than three months old.

Services have changed

New trades, discontinued areas, a second location, a new team: if the website still shows the business of three years ago, it works against your own positioning. Which pages are needed at all is clarified in the article on the page inventory of a business website.

Competitors visible, you are not

This point only counts if the search used the service plus the location, not your own company name. How visibility in the immediate area comes about at all is set out in the article on local visibility.

How to count honestly

Two rules help. First, only count what an outsider would confirm — so let somebody who does not know the business take a look. Second, each line counts at most once: if you have no access to the system and therefore cannot change anything, tick both lines, but each only once. They describe different consequences, one of ownership and one of being able to act.

The scoring scale: maintain, revise or rebuild

The sum of the tallies produces a recommendation. It is not a formula but a rule of thumb: beyond a certain number of simultaneous building sites, patchwork becomes more expensive than a clean fresh start, because each correction drags the next one along with it (project experience).

TallyAssessmentSensible next stepTime horizon
0 to 2The website carries. Individual updates are missing.Set a maintenance routine: who changes what, at what interval, with which access.Ongoing, a few hours per quarter
3 to 5The foundation is sound, individual areas are outdated.Targeted revision: rebuild the affected pages, keep the rest.One to two months
6 to 8Several layers are affected at the same time.Plan a rebuild. Patching usually costs more than rebuilding in this situation.Two to three months
9 to 10The website is working against the business.Rebuild, after clarifying domain, access and content first.Start immediately, implement in stages

The reason for the threshold at six is simple: up to that point, problems can usually be solved one at a time — rewrite a page, swap a photo, add a paragraph. Beyond it, they interlock. Updating content requires access. Getting access requires the old provider. Making the site usable on a phone requires a different underlying structure. In that situation every single task is blocked by the previous one, and the sum of the partial invoices exceeds the price of a rebuild.

Three points that count regardless of the score

Whether the final tally is two or eight, three findings have to be resolved immediately. Missing access to domain and hosting, because otherwise nobody can act in an emergency. Missing or outdated legal pages, because that creates a liability risk. And a provider who no longer replies, because every further delay makes reconstruction more expensive. These three points are not a matter of economics but of being able to act at all.

A side effect of the scale

Anyone who fills in the test once has incidentally produced the basis for every quotation meeting. Instead of „we could do with something new", there is a list that translates point by point into tasks. That shortens the clarification loops considerably and makes quotes comparable in the first place.

What a website realistically delivers

Before the investment, a sober clarification of expectations pays off. A website is not a sales channel that creates demand out of nothing. It is the point at which existing interest turns into an action — and there its contribution is measurable. 85 percent (Bitkom) of trade businesses now offer at least one digital service, most frequently sending quotes digitally at 68 percent (Bitkom), followed by digital invoicing at 62 percent (Bitkom) and online appointment booking at 48 percent (Bitkom). At the same time 85 percent (Bitkom) of businesses observe a clear customer expectation of constant availability.

Collecting enquiries, at night too

The most common and most easily verifiable contribution: a reachable page with a clear contact route takes in enquiries while nobody in the business can pick up the phone. The precondition is that the path to an enquiry is short and usable on a phone.

Preparing appointments

A page cannot replace appointments, but it can prepare them: what is needed, what it roughly costs, when the business can be reached. That shortens the first conversation and reduces the number of appointments that end without a result.

Prompting applications

83 percent (Bitkom) of trade businesses report a shortage of apprentices, 75 percent (Bitkom) a shortage of skilled workers, and 80 percent (Bitkom) use digital channels to address young talent specifically. How a solid careers page is built fills a chapter of its own.

Taking load off the phone

Opening hours, directions, price ranges, responsibilities: every one of these questions answered on the page is asked less often on the phone. The effect appears in no statistic, but it is the quickest to be felt in daily operations (project experience).

It is striking how much customer expectations have shifted. 89 percent (Bitkom) of businesses report that individual offers are expected, 87 percent (Bitkom) observe markedly higher price sensitivity, and 74 percent (Bitkom) say customers have become more critical. A website that leaves prices, scope of service and responsibility open works against that expectation. How far to go on this is covered in the article on showing prices on your website.

89 percent of businesses report that customers increasingly expect individual offers. At the same time, 87 percent observe markedly higher price sensitivity.

Bitkom, study report on the digitalisation of the skilled trades (2025)

And what it does not deliver

The counter-list matters just as much. Inflated expectations are a frequent reason why website projects are later regarded as a disappointment; an unclear brief is an even more frequent one. The following points are explicitly not part of what a new website promises.

  • It does not create demand that does not exist in the market. Where nobody is searching, even a good page does not help.
  • It does not enforce prices. Anyone who is too expensive for the market stays too expensive with a better presentation.
  • It does not replace availability. A form left unanswered for three days costs more trust than the page can build.
  • It does not deliver instant positions in search results. New addresses need time, and positions cannot be promised.
  • It does not replace a recommendation. In local markets, a personal referral remains the strongest route in.
  • It does not repair internal processes. An enquiry with nobody responsible for it stays unanswered — before and after.

One special case deserves a mention: if the website exists, is current and usable and still nobody arrives, the problem usually lies not with the website itself but with discoverability. That is a separate diagnosis with its own sequence, described in detail in the article on why a website is not found. A rebuild does not automatically solve that problem along the way.

The honest sentence for the first meeting

„We expect X enquiries per month through the site, and we will measure that against Y." Anyone able to formulate that sentence before the project has a project. Anyone unable to formulate it has a wish — and a wish cannot be evaluated afterwards.

The maths: how many extra orders it takes

Now for the number. The question is not what a website costs, but how much additional business it has to carry in order to pay off. Three figures are needed, and every business knows them: the annual cost of the website, the average order value, and the share left after direct costs. Revenue is the wrong measure here — the calculation runs on contribution margin.

payback.txt
Annual cost of the website
  = one-off investment / useful life in years
  + running costs per year

Contribution margin per order
  = average order value x contribution margin rate

Additional orders needed per year
  = annual cost / contribution margin per order

Worked example:
  Investment 3,600 euros, useful life 4 years   ->   900 euros
  running costs (hosting, domain, upkeep)       ->   900 euros
  annual cost                                   -> 1,800 euros

  Order value 1,200 euros, margin 30 %          ->   360 euros
  1,800 / 360                                   ->     5 orders per year
Average order valueContribution margin at 30 percentAdditional orders needed per yearWhich equals
400 euros120 euros15roughly one order per month
1,200 euros360 euros5one order every ten weeks
4,000 euros1,200 euros2two orders per year
12,000 euros3,600 euros1one order per year

The table is calculated with an annual cost of 1,800 euros. If your figures differ, put them in — the statement stays the same: the higher the average order value, the faster the investment pays for itself. In businesses with four-figure order values, a single-digit number of additional orders per year is usually enough. For small-ticket services the threshold is higher, but the overall number of enquiries there is larger.

Why contribution margin rather than revenue

Calculating with revenue makes you look richer than you are. An order worth 1,200 euros does not cover 1,200 euros of costs, only the part left after materials, subcontracting and working time. If you do not know your rate, use a conservative value for the rough calculation and check it at the next annual review. In this calculation, an over-cautious assumption is the cheaper mistake.

The economic backdrop argues neither for hectic action nor for waiting. In the German skilled trades, 1,038,126 (ZDH) businesses were most recently registered, generating around 783.2 billion euros (ZDH) in revenue excluding VAT in 2025. At the same time, real revenue in the licensed skilled trades in the first quarter of 2026 was 2.1 percent (Federal Statistical Office of Germany) below the same quarter of the previous year, and in structural construction as much as 7.3 percent (Federal Statistical Office of Germany) below. In that environment, an investment that pays for itself with two to five additional orders per year is a manageable decision — and one that can be worked out beforehand. How to separate one-off and running costs cleanly is shown in the article on budgeting for website projects.

Rebuild or patch up: the practical difference

Both routes are legitimate; they differ mainly in how they treat what already exists. Patching up means the underlying structure stays and individual parts are replaced. Rebuilding means the underlying structure is replaced and usable content is carried over. The second route is not automatically the more expensive one, because a large part of the work is identical in both cases — writing texts, sorting photos, defining structure.

CriterionArgues for patching upArgues for a rebuild
State of the contentTexts and photos are current and usableContent is outdated and would have to be redone anyway
Usability on a phoneWorks, only individual spots are awkwardFundamentally unusable, zooming required
Access and ability to actAccess exists, changes are possibleNo access, no contact person, no documentation
Legal statusLegal pages present and currentLegal pages missing or not matching how the site operates
Scope of changesOne to three pages affectedStructure, navigation and page inventory affected
Time perspectiveThe site should last another one to two yearsThe solution should carry the next four to five years

Two topics explicitly do not belong in this decision, because they only arise afterwards. One is redirects: if addresses disappear or change during the switch, every old address has to point to its new counterpart, otherwise hard-won visibility is lost. That is a craft step with a fixed sequence, described in the article on a relaunch without ranking loss. The other is diagnosing missing visibility, which has its own sequence in the article on the causes. Both topics matter — but they do not answer the question of whether the step is worth taking at all.

  1. Count: fill in the self-test and note the total.
  2. Clarify: secure access to domain and hosting, regardless of the result.
  3. Calculate: put in annual cost, order value and contribution margin.
  4. Compare: put two or three routes side by side before requesting quotes.
  5. Define: formulate a measurable expectation the result can be checked against later.
  6. Implement: content first, technology second — not the other way round.

What else helps before deciding

Between the findings and the signature there are three steps that cost nothing and make the decision considerably safer. In order, they answer: what does a finished result look like, what may it cost, and which route suits the business.

Look at a finished result

Before spending money, it helps to look at a finished page for a comparable type of business: how the path to an enquiry is built, where prices sit, how the page behaves on a phone. The demos for different types of business show exactly that.

Set the budget beforehand

One-off investment, running costs and useful life belong in separate columns, otherwise every figure looks either too high or too low. The article on one-off and running costs provides the structure for it.

Put the routes side by side

Building it yourself, having it built, or working with AI support: the routes differ in effort, commitment and follow-up costs. The comparison of routes to a website sorts out the differences without judgement.

Anyone who has taken these three steps then has a different conversation. Instead of taste, the findings are discussed; instead of lump sums, annual costs; and instead of possibilities, an expectation that can be verified afterwards. What such a sequence looks like from the first sketch to the published page is described in the overview of the route from idea to finished website; the classification of the different approaches is set out in the overview of the differences.

The short version

Count the ten triggers honestly. At zero to two tallies a maintenance routine is enough. At three to five, a targeted revision of the affected areas is worthwhile. From six upwards a rebuild is usually cheaper than patchwork, because the building sites block one another. Regardless of the total, access to domain and hosting, current legal pages and a reachable contact person have to be resolved immediately. And work out beforehand how many additional orders the investment has to carry — with four-figure order values it is usually fewer than expected.

Sources and studies

This article is based on data from: the survey by the Federal Statistical Office of Germany (Destatis) on the use of information and communication technologies in enterprises, in particular the share of enterprises with their own website by size class; the press release of the Federal Statistical Office on revenue in the licensed skilled trades in the first quarter of 2026; the Bitkom study report on the digitalisation of the skilled trades, based on a representatively weighted telephone survey of 504 trade businesses in Germany conducted in 2025; the key figures of the German Confederation of Skilled Crafts (ZDH) for 2025; the press release of the Bavarian State Office for Statistics on internet use by Bavarian enterprises; the Think with Google analysis of mobile loading times and bounces; the performance and accessibility chapters of the HTTP Archive Web Almanac 2025; the Aktion Mensch test report on the accessibility of German online offerings; and the German Accessibility Strengthening Act (BFSG) as published on Gesetze im Internet by the Federal Ministry of Justice.