If you landed here directly, check out the first three posts of this series of - The complete industrial designer:
Stage 1: Research & observation
Stage 2: Sketching
Stage 3: Form & proportion
Clear bottle is actually a harder brief than it looks. A clear body shows every scratch, because there's no texture and no colour to hide and light goes through the damage. It shows flow lines and weld lines from moulding, which an opaque part hides better. It will also show fingerprints and even whatever is growing inside it, which is the point, and also means if there's bad cleaning, it'll show. And a single-wall clear bottle full of cold water will create damp surface, which wets a bag and rings a desk.
I'm okay with this because a clear bottle shows you the water level so you drink, and it shows you the film on the inside wall so you wash it. The clear body decision is actually doing the job of the bottle.
Material selection
First, on materials. I learned about Tritan from Xiaomi bottles. That's where I'd start. It's a co-polyester, genuinely clear, tough enough to survive a drop, it's food-safe so BPA is sorted. It seems like a proven, benchmarked material to begin with. PP is also considerable but I haven't seen one which looks shiny and polished like glass after years of use. Glass is another one but it breaks and I haven't been able to be comfortable carrying a glass bottle anywhere.
Let's also bake in some learnings from Apple. A colleague (ex-Apple) shared that Apple spent years with material scientists on anodising chemistry and ceramics and milling aluminium. Basically they didn't pick from a supplier catalogue and made the materials. We don't have to go to that extreme.
The version of that here could be a hard coat on the clear body. A scratch-resistant coating on Tritan could make our bottle look two years from now. If we do this, then that's something our competitor can't buy off a shelf. Since most CMF work is selection from a catalogue, a minor development becomes the difference.
Colour selection
Second comes colour. We've all learned that the CMF you do at college and the CMF a factory can produce have a huge gap. Oh, by the way, I had written early versions of CMF field guides here & how to build CMF as a capability. This is during the time I took an online course last year.
Take this as an opportunity to document and standardise CMF. Objective is to rule-out subjectivity from within the ID team too. So, what I'd do is write down the L*a*b* values off a physical master sample. You'd start with a Pantone shade card, yes but it should not be the final hand-off. The final handoff should be spectrophotometer reading in L*a*b*, tolerances, a master chip, a signed limit set, and the lighting you're comparing under.
A quick explainer though before we go deeper. L is lightness (0 = pure black, 100 = pure white). a* is red-green axis (positive value indicate red; negative value indicate green). b* is yellow-blue axis (positive value indicate yellow; negative value indicate blue). One this is set, the next one is important - Delta E. Delta E (ΔE) calculates the distance between the master colour standard and the production sample numerically. This value under 1.0 is where the difference is imperceptible to the human eye. 1.0 - 2.0 is visible only to the trained folks with close inspection. And so on. You should have a good grasp on these.
Orange is the hard case here. By the way, I'm using orange colour as a frame of reference from my Stanley bottle and not necessarily what we end up going with. Saturated oranges fade faster and then they immediately look dull. They fade faster under UV so any colour in this territory should have a fade test included before you commit to colour options.
Once done, create three physical parts which would be your limit set. Master, lightest acceptable, darkest acceptable, signed and dated, one copy with us and one at the supplier & factory. That's the single thing that ends colour arguments and repeated verification.
Then the lighting, so D65 daylight for the standard plus a check under warm indoor light, because a bright orange goes noticeably muddier under a 3000K kitchen light and Indian homes are full of 3000K. So we should have a light box studio setup for all colour inspection to mimic both D65 and kitchen/ home warm light.
Finish
I'd put arguments on the table first. A texture anywhere on the clear body is mostly a bad idea. It kills the transparency locally and over time it holds grease and with years of wear & tear, the locally worn off textures will have a bad shiny patch. If a grip texture is needed then it belongs on the silicone, which is where texture actually helps.
The clear body with a polish rather than a texture is what I'd go for., and on a moulded clear part the polish is on the tool. That means SPI A-1 or A-2 diamond finish, which is expensive, slow, and shows every defect underneath it. The hard coat should come in here as post-processing.
The silicone sleeve should have a matte finish, because glossy silicone looks oily and picks up fingerprints easily.
Other parts
The sleeve (if highlighted accent is in the body) could be silicone, and platinum-cured rather than peroxide-cured, because peroxide is cheaper and off-gasses a rubbery smell. If it's silicone, define the Shore value ranges here as well. Similar for the seals esp. on the decisions of keeping it harder for more force on the latch or software so sealing is more comfortable but less assuring.
The cap body will be a tricky one. Tritan vs. glass-filled vs. PP are the obvious choices at first. I'm not keeping ABS right now to offset Tritan body cost. Tritan would match the bottle and scratches uniformly so won't look out of place. Glass-filled may be able to hold the hinge and a latch far better and is opaque but feels flat. PP is cheap and works great with hinges but goes dull almost immediately.
I'd probably run the cap in something opaque and structural and let it be the white part of the identity, so the clear body carries the water and the cap carries the mechanism.
What it looks like in two years
Make every decision based on how the bottle would look like in two years rather than on the day of production. What does it rub against and how often, so the finger zone and the base and the inside of a bag come into your testing criteria. What does it get near, so cooking oil in the air and lemon and coffee and sanitiser and sunscreen residue on palm when leaving for work. And what yellows. Clear polymers go yellow under UV, so a bottle that lives on a car dashboard or a windowsill is a harsh criteria to test against. You should go and put these as inputs to resin selection.
Write clear specifications
Couple of pages with some images and lots of technical annotations.
Body in Tritan, tool polished to SPI A-1, with a hard coat specified and a supplier named, and the reason written next to it.
Cap in an opaque structural polymer, named, with the polymer count for the whole cap written at the bottom and a target of two including the seal.
Sleeve in platinum-cured silicone with specified shore values, matte value, and fade test.
Colour as L*a*b* with ΔE under 1.0 between the sleeve and any coloured cap part, checked under D65 and under 3000K, with a signed limit set of three physical parts. Much later in the project, keep a colour chip and a moulded sample photographed under a kitchen light and under daylight, side by side.
Seals in platinum-cured silicone in a colour you can see.
By the end of this activity, your CMF documents shouldn't look like a deck of collages. Collages are how you decided. That through the final specification sheet is the final work delivered.