From pizza dough to consulting: why make-or-buy is a craft question

Saturday evening. I am kneading pizza dough. Again. Three days of cold fermentation in the fridge, 65 percent hydration. A flour mix of durum with a touch of Manitoba. The twenty-fifth time this year.

The first time, ten years ago, was frustrating. The dough refused to develop, the crust turned into cake. By the hundredth attempt I could tell from the steam above the pan whether the stone had reached 380 degrees or was still at 320.

What this has to do with consulting is the question that occupied me this week.

Buying pizza or making pizza — both are valid

I have not bought a pizza dough in ten years. That is not a recommendation, just a biographical fact. Two paths are open in front of every pizza lover, and both work. If you have a good pizzaiolo nearby, you spend twelve euros and get a result I cannot reproduce at home, even with a domestic wood oven. If you buy a frozen pizza for 2.99 euros, you get something that fills you up, and sometimes that is exactly the right call.

Both paths have a quality anchor. People who buy rely on the producer’s expertise. For the pizzaiolo: possibly an AVPN certification — the Associazione Verace Pizza Napoletana was founded in Naples in 1984, and Neapolitan pizza has been on the UNESCO list of intangible cultural heritage since 2017. For the frozen producer: an industrial process with defined tolerances that reliably lands in your local supermarket.

People who make rely on their own craft. On repeated practice and on the correction that follows the last failure. That path is slower, more frustrating and almost always worse than buying at the start. Around attempt one hundred, the relationship flips.

Both are valid. But the two paths require different evaluation yardsticks.

What a pizza school actually requires

To earn a pizzaiolo diploma in Switzerland, candidates work through three certification levels, prove at least one year of professional experience, and complete a minimum of 80 hours of formal study, according to Pizzaioli.ch. Then comes the practical exam. Only then the diploma.

That is not arbitrary effort. It is the threshold from which a human masters the craft. Below that threshold, the person is an apprentice — an honest and useful position, but not a pizzaiolo.

Now compare that with the job title “consultant.” There is no diploma, no binding curriculum, no threshold. Anyone who writes “ERP consultant” on their business card tomorrow may do so. The same applies to “AI strategy consultant.” The market consequence: massive quality variance, with buyers operating without a yardstick.

The 10,000-hour trap

Malcolm Gladwell popularised the 10,000-hour rule in 2008. Spend ten thousand hours on a discipline, the story went, and you become an expert. Catchy, but sloppy. The researcher behind the original study, K. Anders Ericsson (NCBI review 2019), rejected Gladwell’s reading himself. The 10,000-hour average came from twenty-year-old violin students who were still nowhere near mastery at that point.

Brooke Macnamara’s 2014 meta-analysis settled the empirical question. Deliberate practice explains roughly twelve percent of performance variance on average. Twenty-six percent in chess, twenty-one in music, eighteen in sport. The remainder lies elsewhere: quality of coaching, feedback, individual fit, domain specifics.

For evaluating consultants, that means something concrete. “Fifteen years of experience” on a website proves nothing. Someone who has not corrected the same mistake across fifteen years has not gathered practice; they have repeated the same day. The right question to ask a consultant is not “how long have you been doing this?” but “which decision did you reverse in a comparable project, and what did it teach you?”

An honest answer to that question reveals whether someone holds the craft in their head, or only on their business card.

Make-or-buy as a craft question

The PwC Mittelstand IT outsourcing study from 2024 reports that 62 percent of the German Mittelstand has outsourced parts of IT. Common fields: infrastructure, helpdesk, IT security. Over 70 percent of respondents confirm higher availability and lower downtime.

This is the make-or-buy statistic that appears in every other consulting deck. What it does not answer: how many of those 62 percent could have rejected the wrong provider? How many had the yardstick to tell craft from marketing, before the contract was signed?

My estimate is well below 62 percent. That is not disrespectful. It is the awkward property of the consulting and outsourcing market: the buyer must have evaluation competence before purchasing. Without it, the buyer does not buy a result. The buyer buys trust. And trust is always the most expensive line item on the invoice.

The Schneider & Wulf practitioner guide from March 2025 names four pillars for a sound make-or-buy decision: cost, competence, time, strategy. That is useful but incomplete. The fifth pillar, often forgotten, is evaluation competence in-house. Whoever outsources must be able to evaluate the provider, or they are buying blind.

And evaluation competence cannot be sourced. It has to be built.

The two lobbies that miss the point

This is where the debate becomes polarised. Two loud voices fill the room, and both miss the actual lever.

The insourcing lobby says: “Do everything yourself. Own skills, own people, no vendor lock-in.” That sounds like sovereignty. It overlooks that internal skill-building can take a decade, and the market shifts during that time. Anyone who started building SAP skills in 2015 is now in an S/4HANA world with cloud-native components nobody learned during their apprenticeship. Full insourcing is romantic, rarely realistic.

The vendor lobby says the opposite: “Let us handle it, you focus on your core business.” That sounds like efficiency. It is dangerous, because evaluation competence gets outsourced along with everything else. A buyer without a yardstick cannot lead the provider. The provider becomes the implicit strategy function of the company, a role they were never hired for.

Both lobbies miss the point. The real lever sits in between. The company retains the evaluation function in-house, even when execution runs externally. The company builds enough skill that the managing director does not need to become a pizzaiolo, but can recognise one.

That is also the trap too many Mittelstand companies fall into when picking the wrong project manager — they delegate evaluation to someone who has not yet developed a yardstick.

My own path — analytical, slow, hard

Here it becomes personal, because Pillar C demands it. I work analytically and I take the hard path. That is not a virtue, just an observation about my work that has become clearer over recent years.

Concretely: a first conversation with me is not a pitch. It is a test of the managing director’s evaluation competence. I ask back. I challenge assumptions. I name the cost of my sparring session, and the cost another consultant would quote for the same outcome if they immediately offered “we’ll handle it for you.”

The fast path would be to send a proposal. More mandates, faster revenue. The hard path is to put the client in a position to evaluate me before buying. That does not happen in twenty minutes. It happens in a sparring session that takes several hours, which I give away, because I would rather land the right mandate later than the wrong one now.

This stance has an economic consequence that is not pretty. I sell fewer days than a consultant who turns every lead into an engagement. What I sell tends to last. Follow-up mandates come without a new pitch. Referrals come from mandates that have ended, not from rolling contract extensions.

That is the hard path. Anyone who walks it lives with the fact that it does not scale like a typical consulting model. Just as a pizzaiolo at a wood oven does not produce one hundred pizzas per hour. Not a weakness, just the architecture of craft.

If you want to understand what preparation a project needs before the first consulting day, or why internal skills and AI adoption belong together, start there. The sparring session comes afterwards.

FAQ

How can a managing director recognise a good ERP or AI consultant without deep personal expertise?

You do not need to become a consultant. But you need enough yardstick to place three answers in context. How would the provider describe the worst day in their most recent project? Which concrete decision did they reverse in a comparable project, and why? What would they do differently today than two years ago? Clean answers to all three are a reliable craft signal, dodges or generalisations a reliable marketing signal.

Is it worth building IT or ERP skills internally, or better to outsource permanently?

The PwC 2024 study says 62 percent of the Mittelstand has outsourced IT parts. That is observation, not recommendation. The real question is: which skills must stay in-house so that you can lead an external provider? At minimum, evaluation capability. That is often outsourced along with everything else, and that is precisely the mistake.

How long does it take for an internal employee to become genuinely ERP-competent?

The pizzaiolo answer is honest. One year of professional experience combined with 80 hours of formal study and several certification levels, and then real learning begins. ERP is not faster. Anyone aiming to go from zero to productive in two years either hires a senior or lives with apprentice-level output. Both are valid as long as you call them what they are. Selling apprentice output as senior-level is self-deception.

What is the difference between a consultant’s hourly rate and “consulting in-house”?

The hourly rate is more transparent but rarely cheaper. A good external consultant brings the yardstick knowledge from 30 projects. An internal employee brings the context knowledge from your house. Make-or-buy asks the wrong question. The right one: which share of those two knowledge layers am I missing right now, and where can I source it most cheaply?

Does training own employees beat buying consulting?

Both. Training builds long-term craft. Consulting closes specific gaps quickly. Neither replaces the other. A pure-training strategy gives you a good team in five years and a compliance problem in two. A pure-consulting strategy gives you a finished project in two years and no internal shoulders in five.

Why is the analytical, hard path better than the fast one?

Because it is reproducible. The fast path sells “we will handle it for you”. The analytical path sells “we will show you how to recognise it, and decide together whether we handle it or you do”. Across a project lifecycle, the first sells more days, the second fewer days and more follow-up mandates without a fresh pitch.

Next step

If you are facing a make-or-buy decision (ERP, AI, migration, data audit), the conversation does not start with a proposal. It starts with a free 60-minute sparring session. You bring the concrete question. I bring the evaluation logic. The outcome is your call.

My main services outline the topic areas I work in. The about page helps you decide whether the sparring style fits your expectation. Direct contact: rene@pfisterer.xyz, or via the contact form.

I will still make the Sunday pizza myself.