I usually handle the counter layout for our games - I can count the exceptions on one hand - but, NATO symbols aside, it's usually someone else doing the hard stuff, drawing the soldier or the horsey or the cannons. Each of those illustrations naturally increases the art budget and pushes our break-even point further down the line.
One of the appeals of doing a series like Shields & Swords II is that we could pay for unit illustrations one time and then reuse them in subsequent volumes. In fact, even though the first release, The Grunwald Swords, only used five different unit types, we commissioned the art for three more units, knowing we would use them later. One of them, the crossbowman, still hasn't shown up in an actual game, but we'll get there eventually.
We knew going in however that the Shields & Swords Ancients offshoot wouldn't be following the same model; each volume would need its own set of one-off illustrations. This would increase the budget and the break-even point, and concomitantly the risk. I also wanted something that looked and felt quite different than the medieval games in order to distinguish the two series, and I was having trouble verbalizing what it was that I wanted, which made it difficult to choose and approach an illustrator.
Meanwhile, we invested in a Wacom tablet to help me step up my graphic design game. I tried my hand at doing counter illustrations for a few games, such as the already released Antony and Cleopatra and the forthcoming District Commander series and At All Costs!, and I was very pleased with the results. Don't get me wrong: I'm not really an "illustrator" so much as an amateur graphic designer with pretensions. They're more like icons than figure drawings, relying heavily on silhouette, though the handful of character headshots are actually pretty competent. But I did well enough with these that I felt like I could probably do the icons for With It Or On It myself.
Given my predilection for silhouettes, it's no surprise that I gravitated toward black figure pottery as my primary source of inspiration. Strictly speaking, black figure saw a steep decline in popularity around the time that my game starts, having been supplanted by the dominant red figure style. I justified the seeming anachronism of modeling my icons in the black figure style for two reasons. First, it wasn't really as anachronistic as all that; there are black figure vases that date from the second century BCE, long after they were out of vogue. That's not so surprising; westerns and musicals went out of style in the sixties, but it still seems like every year there's at least one of each coming out at the cinema. Secondly, I liked the black figure style better than the red figure, and it was my game.
Two of the illustrations - the hoplite and that goofy-looking horse - were sourced directly from extant black figure pottery. That is, I based my illustration more-or-less on the historical artwork. For the two others, I didn't have black pottery sources, and so tried my best to ape the look - including the naïve anatomical proportions.
While these counters weren't used in my very first prototype, I had a version of them in my second. What surprised me was how doing the art also influenced the development of the game itself. The short version is that Units have a Fresh side (no stripe) and an Exhausted side (stripe), with only Fresh Units able to move or attack. Naturally because of this Exhausted Units did not have a combat class designation. This is different from the medieval series, where Units that take step losses are flipped to a reduced side that sometimes (but not always) results in a reduction of their Combat Class. I felt the absence of this variability and unpredictability acutely, but wasn't sure how to fix it.
So, I'm laying out the playtest counters with my new illustrations, and I know I want a stripe for Exhausted Units but I don't know quite where I want it. I experiment with a couple of different options, each on their own Photoshop layer, trying to see what looks best with these illustrations. This is a step that would normally come at the end of the process - after the game is absolutely positively finished and we hire an artist to do some illustrations. While I'm tooling around with it, I happen to leave both layers visible, and now I have two stripes. And almost immediately I start to think about what might be represented by two stripes, and come up with the brittle unit mechanism, which suddenly interjects an incredible amount of tension and uncertainty into the proceedings. That's a major design problem solved, and that solution was inextricably linked with doing the art and finalizing the counter layout in that early stage.
I remember reading something another designer wrote - can't remember the designer and can't remember the game, but I remember that they did all the illustrations and graphic design for it. They said that the two things, the game design and the presentation, were so inextricably linked that they couldn't imagine doing one and not the other. I thought that was kinda nuts at the time, and goodness knows that I could never be trusted to do a mapsheet, but I kinda get it now.