Guide · Layer 1

Draw the Process

Written for: seed / Series A scientist-founder

Zoom in on your core technology

Draw just your core technology as a set of boxes & arrows:

what goes in
what transformation happens
what comes out
the conditions it needs

Early on a mix of words, numbers, and a sketch in Powerpoint is great. If a metric is what makes your tech groundbreaking (e.g., some yield, efficiency, selectivity), it must have a home here!

🧭 Coach’s Read

The deployed version of your tech usually isn’t the lab version. If something you depend on isn’t something you can buy in bulk, it should go on the diagram / be accounted for.

Zoom out to the full process

Now back up to the whole process. A few high-level scoping decisions to iron out before you fill in the flow:

Reference process: find the closest existing process and use its flowsheet as your template
Unit of analysis: work back from how you’ll sell the product or service, and assume a simple business model for now
System boundary: where the flow starts (inputs you can buy) and stops (a product people actually buy)
Scale: a representative, at-scale footprint the industry actually builds at

🧭 Coach’s Read

End the flow at something people actually buy. Unbaked cakes have no market — if buyers want 99.5% purity and you stop at 97%, you have to internalize that finishing step if you want to quote the price of 99.5% purity in your model (someone has to do the work, but then also then get paid for it!)

Fill in the boxes

Now draw the rest of the process, one box per unit operation or process step:

Cover the whole route first: every major unit operation gets a box, so the total cost you land on reflects real production
Stay at one altitude: like walking a brewery and noting each big piece of equipment, vs opening up some here and there or taking apart others
Use references to find blind spots: comps, Google, even an AI chatbot for a feel of how the industry runs; you’re checking that no big operation is missing
Find comps, don’t average them: look up a few real processes, then pick one representative flowsheet and build it out end to end; averaging several together gives you a plant that doesn’t exist anywhere

🧭 Coach’s Read

Get end-to-end first (I’d rather see a rough box for every step than a ton of detail that also slows you down). You also can’t tell which box matters until you look back at the whole process with some analysis.

Plug in your core technology

Take the core tech you drew above and map where it slots into the reference process; that gives you your future process flow.

A simple swap: drops into an existing step, like a rechargeable AAA for a single-use one in your TV remote
A real integration: forces changes around it, like an AA (which is wider) that needs more space than the AAA it replaces

When it’s a real integration, treat it as its own small piece of work: figure out what changes, by how much, and where the new boxes go, then talk to people who’ve run something similar.

🧭 Coach’s Read

The integration costs are the ones teams forget. If your tech needs new couplers, conditions, waste treatment, etc - these can somestimes come back to bite you!

Next: 1.3 — Mark the delta & get feedback →