Start with the verdict. In the exploration phase a sketch has one job — to make a decision, not to present a drawing. A sketch that moves a project forward has to answer a specific question: is this volume relationship right, which way does the part line run, does this surface need to tighten? A render answers "does it look good"; a sketch answers "why this and not that." Three kinds of sketch do the work: small volume studies that fix proportion, surface studies that fix the part line, and cross-sections that fix ergonomics. There are not many of them, and every one needs a reason to exist.
The short version: work in the order volume → surfaces → ergonomics, and solve exactly one question per stage. Volume: thirty 30-second sketches on A4, outline and visual weight only. Surfaces: part lines and transitions, one line per sheet. Ergonomics: cross-section plus contact points, checked against real dimensions. Do not start with detail, and do not start with colour.
The 12 polished renders that got sent back
Last year we were sketching a handheld inspection device. Round one, I handed over 12 finished marker renders. The client looked at all of them and said: "Let me see something else." Round two I handed over 18 thirty-second sketches, and the direction was fixed on the spot. That one stung, and it taught me something useful: the client was not choosing between drawings, he was choosing between directions. In round one I had not given him a direction — I had given him one beautiful conclusion.
Round one: volume only, 30 seconds each, thirty sheets
For the first two hours after a brief lands, I do not touch detail. A4 cut in half, one 0.5 mm gel pen, one sheet per 30 seconds, thirty sheets. The rule: outline plus a single main division, no material indication, no shading.
Why the 30-second limit? Because a slow hand starts decorating, and decoration pulls you into detail, and detail kills the volume judgement. Thirty seconds is just long enough to record a sense of weight.
Three things are actually being compared at this stage: the proportion between length, width and height; where the visual centre of gravity sits; whether there is one feature anyone will remember. For the inspection device the brief said "professional, but not cold." Professional came from a narrow, elongated proportion with the mass low; not-cold came from generous radii and a three-part division. Both of those have to be visible at the volume stage. Adding them at the colour stage is too late.
Our internal test for a volume sketch is one question: shrink it to two centimetres square — can you still tell which product it is? If not, the recognisability is being carried entirely by detail, and the volume itself is empty.
Round two: surfaces and transitions, one line per sheet
Once the direction is fixed the question changes from "what does it look like" to "how does this surface transition." Now I draw zoomed details, one line per sheet.
Three types:
Part-line sketches — mark at what height the line wraps around, and how it clears the buttons and ports. A part line that drifts immediately costs a grade, because it decides where the shadow falls.
Transition sketches — flat-to-curved, or hard edge with a small fillet. The same 2 mm radius reads soft on a small product and slack on a 300 mm body. That has to be drawn side by side to be argued about.
CMF zoning sketches — which area stays gloss, which goes matte, where the colour breaks. No colour at this stage; hatching density carries the light-dark difference. Dense hatching usually becomes matte, the untouched areas gloss.
A mistake we paid for: a small desktop appliance where, at the surface stage, I drew the top-to-side transition as one continuous radius. It looked fine on the render. At the prototype stage that arc caught a long bright band under side light and cut the body in half. We changed it to a two-stage transition with a 15 mm flat between. That class of problem is fully predictable at the sketch stage — provided you are drawing the transition, not the outline.
Round three: ergonomics, with a cross-section and a hand
For anything handheld I always finish with two sheets: a horizontal cross-section, and a hand grip study.
The cross-section gets dimensions written straight onto it. For grip width, adult hand breadth is commonly cited at 80-90 mm, and the corresponding grip diameter lands in the 40-55 mm band for comfortable force; past 60 mm one-handed grip gets awkward. These are the usual industry ranges — on a real project the client still has to hold a sample and confirm.
The hand study does not need to be lifelike. Outline the palm and four contact points, and that is enough. What matters is the thumb landing, the index position, and where the button goes. On one product we had put the main button 8 mm above the thumb's natural landing; the moment we held the prototype everybody felt it was wrong. Had the hand been drawn at the sketch stage, that 8 mm would have been caught for free.
Should exploration be capped? Yes. Thirty sketches at the volume stage, six to eight at the surface stage (each with one specific question), two to three at the ergonomics stage. Forty sheets in total, and the approval rate is higher than five tight renders. Detailing before the direction is fixed means investing time in work that will be thrown away.
When not to sketch
Two cases where sketching wastes time.
One: the client has already supplied a reference and says "make it feel like this." Now the right move is to photograph the reference and draw corrections on top of the photo. A sketch loses information in the transfer.
Two: the form has already been locked by structure — internal stack-up fills the volume and the shell is down to filling the outline. Then the job is to align structure and appearance boundaries first; sketching can wait until the boundary is clear.
| Stage | What to draw | How many | Question it answers |
|---|---|---|---|
| Volume | Outline + one main division | ~30, 30 s each | Are proportion and visual weight right |
| Surface | Zoomed detail, one line per sheet | 6-8 | Part-line position, hard or soft transition |
| Ergonomics | Cross-section + hand contact points | 2-3, dimensioned | Can it be held, where is the button |
| CMF | Hatching zones, no colour | 2-3 | Where gloss breaks from matte |
Our current practice is that every exploration sketch uses the same pen and the same paper, with no visible difference in finish level. That way the client sees a dozen directions, not the first draft of one artwork. In the years we have been doing this, the projects that went right in one pass were not the ones with the tightest drawings — they were the ones where the directions were laid out most completely.
If you are working on a form right now, start with one exercise: spread out the sketches you have and ask each sheet what specific question it answers. The ones that cannot answer are not exploration. They are decoration.
Written by the Hezi Industrial Design team, from project practice in Dongguan. Hezi does appearance design, structural design and CMF as one service — 320+ products in production. Website: www.hezidesign.com
