Wissen, Dynamik, ÜberlappungKnowledge, dynamics, overlap

Part 4 of 5

Welcome to part four. We stay with the phase model and look at how knowledge, concepts and environment change during the project, and when phases may overlap. ShopNova is back again.

Integration, ignorance, degree of innovation

Detail concepts must fit into the overall concept. So you keep embedding them there in your mind, a successive integration process. That's why the principle is: from rough to detail... and back again. Along the way you may notice that influencing factors have changed, that new interfaces turn up, or that goals are not achievable. Your knowledge of the system grows over the course of the project, but it flattens. At the start it's never zero, at the end never complete. A permissible degree of ignorance remains. For safety-relevant systems it has to be small. You reduce it with safety margins, analyses, simulations and redundancy. In addition, the degree of innovation should fall during the project. The more proven solution elements, the better you can estimate and coordinate the result. New things belong where they bring competitive advantage.

Dynamics of the overall conception

An overall concept is rarely frozen so that afterwards you only develop details. During the detail concepts it is checked and adapted again and again. Triggers come from within, through better insight into problem and solution, and from outside, for example through new technologies, competition, financial situation, markets or laws. This creates a trilemma: overall concepts age, because they rest on the state of knowledge at the time of the decision. The longer you plan, the sooner results are devalued. And the further the project advances, the harder the concept is to change. You can't resolve that. But you can behave wisely: prefer smaller solutions with quick benefit over super-integrated large solutions. Plan in flexibility and options. Provide for partial roll-outs. Don't optimise unproductive details. Make uncertain decisions as late as possible. And freeze the concept late and step by step.

Example: Knowledge, environment, dynamics

An analogy for the last two scenes. You're driving in fog with a satnav. At first you see little. The further you drive, the more you see, but it never gets completely clear. In a traffic jam the satnav recalculates the route. An overall concept ages in the same way: new insights and new circumstances force you to adapt. At ShopNova it turns out after the pilot that customers ask different questions than expected. That's insight from within. On top come new rules for using AI in customer dialogue. That's an influence from outside. That's why the team plans small steps, keeps the language model replaceable and fixes the concept only late. Remember the trilemma with three words: Ageing, Devaluing, Freezing. The concept ages, the planning is devalued, and with progress it freezes.

Overlap and immediate measures

Detail studies, system construction and introduction don't have to run one after the other for the whole solution. Partial solutions can be worked out and implemented overlapping: for a shorter time-to-market, with limited capacity or a tight budget, or to achieve interim benefit early. The limit: you only start system construction when the main study has delivered an overall concept. Sometimes you need mental look-aheads. If a building project could fail on the load-bearing capacity of the ground, you do the test drilling already in the preliminary study. Immediate measures create relief and trust in the team early, but can pre-load later solutions. Recommended are measures with little effort, visible effect, so-called quick wins, and as little prejudgement as possible.

Example: Overlap and immediate measures

An analogy. In a kitchen the starter is already being served while the main course is still cooking. That saves time. But the menu must be set before anyone starts cooking. And a plaster stops the bleeding at once, but doesn't replace surgery. At ShopNova the system build for the returns chatbot is running while the detail study for the next topic is still going. As an immediate measure, the most common answers go onto the website as an FAQ right away. Little effort, visible effect. The condition: the overall concept exists, and the FAQ blocks no later solution. You recognise a quick win by three letters: L, V, O. Low effort, Visible effect, Open for later.

Summary of part 4

To wrap up part four, the key points. Knowledge about the system grows but flattens, and the degree of innovation should fall. Overall concepts age. The antidotes are small steps, flexibility and late commitment. Overlapping is allowed, but never build without an overall concept. Immediate measures should be quick wins. Your mnemonics: Ageing, Devaluing, Freezing. A quick win is low effort, visible, open. On to the last part.