From the first click in the contact form to the running software. Seven clear steps, with timings for your effort and ours - so you know exactly what's coming.
Software projects often feel like a black box. You put money in, eventually something comes out - in between, you don't know much.
With us, every step is visible. At any point you can see where we are, what comes next, roughly how much longer it will take, and what you have to do right now (or don't). Here's the complete process - compact, honest, with all the waiting times and all the effort on both sides.
First contact & intake
15-30 min · one afternoon
Tailored questionnaire
45-90 min · at your pace
Three paths to choose
30-60 min · plus decision
Build phase with test access
2-16 weeks · depending on path
Testing with real data
1-3 weeks · depending on path
Handover to the system
1-2 days · plus 4 weeks of follow-up
Ongoing operation
monthly · for as long as you want
You write to us via the contact form or by email. Within one working day, you get a reply with a link to our portal and a short explanation of what happens next. You register there with your email - no password, we send you a sign-in link whenever you want to come back in.
In the portal you answer four short mandatory questions: What's going wrong today? What's your biggest goal? How technical do you consider yourself? What's your role in the decision? Below that, you see six slots for evidence - you can optionally upload whatever would help us: a screenshot of your most important tool, an Excel list you work in, a sample quote. If you prefer speaking to typing, you can also record a voice note. All of it is optional.
Not sure what we mean by a question? Write to hallo@wendwerk.de - we reply within one working day, without you needing an appointment.
Traditional providers start with a discovery meeting - 60 or 90 minutes together at a table or on a video call. That feels personal, but has three drawbacks: you have to find an appointment (often only two weeks out), you speak spontaneously about things you should actually have thought through, and in the end you have no written record - only the consultant's memory of it.
Asynchronous is objectively better here. You write when it suits, calmly, with evidence to hand. We get material we can work with, instead of a phone transcript. The Mom Test (Rob Fitzpatrick, 2013), the classic on customer discovery, shows: good questions plus artefacts beat any spontaneous conversation. Our intake is built exactly along those lines.
After the intake, we prepare a questionnaire for you that's specifically tailored to your situation. Within one or two working days, you get the link by email. The questionnaire holds between 24 and 30 questions - that sounds like a lot, but each one is concrete and answerable in 1-2 minutes. You can pause at any time and pick up exactly where you left off.
The questionnaire walks you through eleven topic blocks: Vision and desired state. Who works with it daily, where it currently annoys. What data and objects exist. What daily life looks like, where it's currently broken. What scale (users, transactions, inventory). What conditions (mobile, offline, GDPR). What other systems are involved. When the project is a success. What would be a deal-breaker for you. How the software should feel. And at the end: what it's worth to you, what fits budget-wise. Money deliberately comes last - so that value leads the discussion, not the number.
Stuck on a question, or unsure what it means? A short email to hallo@wendwerk.de is enough - we rephrase, explain the background, or skip the question.
A universal questionnaire sent to everyone is a wasted opportunity. A solo freelancer shouldn't have to see questions about team structures. Someone working in a law firm should read "cases" rather than "tickets." So we build the questionnaire individually - with an AI that uses your phase-0 answers as anchor, and a manual check by us before it goes out to you.
The order of topics follows SPIN Selling (Neil Rackham, 1988) - but applied to diagnosis, not to sales. First situation, then problem, then implication, then solution, then value. Studies show: thinking in this order leads to more precise solutions than starting with budget. More on this on Before we build, we figure out what's broken.
Once you've finished the questionnaire, we send you a link to the evaluation within one or two working days. There you first see a problem definition: a short text in which we play back how we've understood your situation. "This is how we heard you." If that fits: one click - and you see the three paths. If not: you write us a sentence about what you actually meant, we revise - no rush, no discussion at the table.
The three paths are called Small, Medium and Large. Each has a clear scope, a one-off setup price, a monthly maintenance rate and a timeframe. We mark one of the three as our recommendation - with a concrete sentence on why we'd pick that one based on your answers. But the decision is yours. Once you've decided, we send you a short contract (one page plus appendix, in readable language) and we begin.
Sitting with the decision and want to talk it through? Write to hallo@wendwerk.de or book a short call - sometimes a 15-minute call clears up more than three emails.
Three options is no accident - it's psychologically grounded. The Asymmetric Dominance Effect (Huber, Payne, Puto, 1982) shows: with two options the decision is hard, with four or more it overwhelms. Three are the sweet spot - they structure the field without burying you.
The split into problem definition first, solutions only after, is also deliberate. If we hand you three paths right away, you discuss solutions - before it's clear whether we even understood the problem correctly. This split costs one extra click, but often saves an entire round of iteration.
Once the contract is signed, we begin. You get access to a test version of your software, which is still empty - we fill it step by step with what we agreed. You can click in at any time, look at the current state, try things out, and leave feedback.
At meaningful milestones we send you a short written update - what's been added, what's coming next. When longer phases come in which little happens visibly (because we're working on the data structure, say), you still get a status, so you never lose sight of where things are. What you don't have to do: fixed appointments, weekly calls, preparing for sprint reviews. If you want a walkthrough or want to talk about something bigger, we're happy to set up a time - but none is required. Depending on the path chosen, the build phase takes between two weeks (Small) and four months (Large). The timeframe is known to both of us from the choice, and it's not a moving target.
Something in the test version unclear, not to your taste, or missing? Write to hallo@wendwerk.de or leave feedback in the portal - we respond within one working day.
Iterative development with early access has been best practice in software since the 1990s - popularised by Scrum and Lean Startup. The core idea: whoever builds in silence for months and presents a huge finished product at the end is highly likely to have accumulated misunderstandings. Whoever grants early and continuous access can catch deviations while they're still cheap to fix - rather than in an expensive big-bang review at the end.
We combine this with a second principle: tools first, AI second. As a rule, the build phase initially consists only of classical software that maps your process digitally. AI comes - if at all - in a second stage, at exactly the places where it actually saves time. More on this on Tools first, AI second.
When the build phase is complete, the system is functionally finished - but not yet ready for daily use. Now comes the step many providers cut short or skip entirely: thorough testing with real data and real workflows. You, or the people who'll actually work with it later, run through typical workflows from start to finish - with your actual master data or a realistic copy of it, not with demo examples.
Concretely, that means three things. First: you run through the main chain - "new order to finished invoice," or whatever your backbone is, in several real variants. Second: we tackle the edge cases together - empty database, aborted saves, two people in the same record at the same time, unusual inputs. Third: in parallel we check what you can't see - performance under load, behaviour during network failure, GDPR compliance, backup restore, security against typical web attacks. What we find here, we fix before the system goes live. This phase has its own time budget - it's built in, not optional.
Finding friction, errors or unclear spots while testing? That's exactly what this phase is for - write in the portal or to hallo@wendwerk.de; the more precise, the faster we can smooth it out.
The technical term is User Acceptance Testing (UAT) - testing the finished system against the real requirements of the future user, with real data and real workflows. UAT has been a standard in serious software projects for decades - but in practice it's often reduced to "a few clicks before handover," because testing is the first thing cut under time pressure. Whoever does that only pushes the risk into the time after going live - where it gets more expensive, more visible, and lands under stress.
Studies on the cost-of-defect curve have shown since the 1970s (Barry Boehm, "Software Engineering Economics," 1981): a bug caught in the test phase costs many times less than the same bug after going live. For web software, the factor sits, depending on the source, at 10x to 100x. Put differently: two weeks of careful testing often save several weeks of emergency fixes in live operation - plus the damage to user trust, which can't be put in numbers.
When the build phase ends, comes the moment when the test version becomes the live version. Concretely: your data moves from the trial run to the final servers, you get your final address (your own domain or a subdomain), and if you're replacing existing software, we help with the data migration. You set the exact moment - some want to switch on a Friday evening so the weekend is a buffer, others prefer the first of the month.
If you or your team want a short induction, we come for that separately - either as a 1-hour video call or as an onboarding video you can watch at your own pace. The software should, however, be built so that you get along without training (that's marker two from our principles). After handover, we're available free for four weeks for any follow-up questions - email, portal, quick replies. In these four weeks, you notice the small things that bother you in daily use, and we smooth them out.
Snagging during migration, on the first live day, or during induction? We're especially reachable in this phase - a short email to hallo@wendwerk.de is enough, and we can often jump in at short notice.
The switch from test to live operation is classically the moment most software projects come apart. Suddenly there's real data in there, real users working with it, real friction points appearing. Time-to-value - how long it takes a new user before they use the system productively - is the critical metric. Onboarding research shows: users who experience a tangible benefit in the first 14 days stay long-term. Those who don't, give up.
The four free weeks of follow-up aren't a marketing trick - they're our insurance against having overlooked small things. In these four weeks, we see the real daily use of the software, not the planned use. What stands out, we adjust - often it's five or six small smoothings that make a big difference.
After the four weeks of follow-up, regular operation begins. You pay a monthly maintenance rate that's included in your chosen path - it covers hosting, backups, security updates and small adjustments. Larger wishes - "Can we add feature X?" - we collect, and we propose to you quarterly what would make sense next, with price and timeframe. You decide what comes and when.
If at some point you say "We can do this ourselves," "It doesn't fit any more," "We want to go elsewhere" - no drama. We hand over the source code, the data, the access. It's your software from the beginning. There are no hidden lock-ins, no shackle contracts, no platform binding. You can cancel month by month. That may look like a risk for us - but it's our strongest discipline. Whoever doesn't have to stay has to be served really well.
A question in daily life, or an idea for an extension? Free in the first four weeks after handover, then included in the maintenance contract - just an email to hallo@wendwerk.de, no ticket system, no support portal login.
Vendor lock-in is the most common complaint of SMEs about established software providers. First everything's lovely, then you notice that the data sits in a format you can't import anywhere else, that the contract runs three years, that a migration costs six figures. Open standards and monthly notice are our answer to that - and they force us to be good every month afresh.
Quarterly development proposals instead of monthly newsletters of feature wishes are also deliberate. Lean Software Development (Mary & Tom Poppendieck, 2003) shows: those who think in small steps build better than those who juggle a huge backlog. We prefer working with two or three large extensions a year, rather than twenty mini-patches.
Transparency, your pace, no surprises. You see everything, you decide everything, you can step out at any time.
We didn't invent this approach to look particularly modern. We built it because the classical version - book an appointment, spec sheet, quote, contract, months of silence, handover workshop, eternal binding - almost always ends in frustration for SMEs. The software in the end wasn't what was thought. No one noticed something was heading in the wrong direction until it was too late. Whoever doesn't want that has to do it differently. That's exactly what these seven steps are for.
More on how we really understand the problem in the first steps is on Before we build, we figure out what's broken. Why we split the build phase into two stages and when AI comes into play, you'll find on Tools first, AI second. The quality standard for everything we build, you'll read on When is software actually good.