Cindy or Naomi? How to know which eComm models will boost ROI using retail analytics

Anna-Louise McDougall
August 29, 2022
9 min

Squeezing a whole collection shoot into two days? Calling in samples, photographers, hair and makeup, and spending your precious budget on 3 pairs of styling shoes worth your pay packet? Whether you shoot twice a day, twice a week, or twice a year, it pays to stack up your model stats.

Product imagery isn’t usually considered the most creative aspect of the brand, but - no matter what kind of retail store you run - it’s what your online customers engage with the most.

It’s the final stop in your efforts to have them click ‘Add to Cart’. Truthfully, there are plenty of opportunities to get creative and not only remain commercial but boost your AOV, amongst other business vitals.

So, before you start staring down the barrel of a talent agency invoice and wondering how you let your budget blow this far, it’s prudent to be clued in on determining the right eCommerce model for your online store. There are often overlooked tweaks you can make to help you easily break even on that agency fee and start scaling.

And while there’s never going to be a one-size-fits-all formula for your model choice, there are steps you can take to execute best practices for your brand.

Here’s how to make your eCommerce product models really werk for your online retail store.  

Choosing your eCommerce model(s)

A good place to start. There is no doubt using models for your eCommerce site is worth every penny for their time, especially if your data tells you that on-model imagery translates into better sales. Plus, there’s not much point in low-balling the fee for the risk of having to reshoot… and there’s only so much retouching you can afford, right?

Ticking these checkboxes should put you in a pretty good position on shoot day:

1. Does the model tell your story?

Every eCommerce platform is selling a unique story to its consumer about who they are, or want to be, after making a purchase.

For example, Net-a-Porter has long served the fashion-forward set with cash to splash using partially headless models, high-end styling, and a typical runway body type.

web screenshot of Net-a-Porter website depicting model wearing Valentino black sweater

The intention is to spark the same ‘I gotta have it’ response that one might get flipping through Vogue.  The difference? The customer can act on the emotion immediately.

Patagonia’s reputation precedes the clothing, so it’s no wonder they play into their outdoorsy tomboy aesthetic, with models who look like you might find them at the base of a mountain. With smiling faces, little to no styling, and natural hair and makeup, the customer is able to discover the product without distraction, which suits the technical, long-wearing nature of the garments.

web screenshot of Patagonia website showing woman wearing mustard yellow Patagonia raincoat

Even though both retail sites lead with ghost or flat imagery, the story is told by showing how the clothing fits and functions on a model best fit to represent the customer, and more importantly, how to wear it.

2. Does your choice match historical data?

Everyone on the online team (and every team) has that one model who just blows their socks off wearing the brand.  But, does that model resonate with your audience as well?

It’s just like when the buyer might want to invest like crazy in particular styles because they’re #obsessed. Will the audience be #obsessed too?

There’s only one way to find out if your users are engaging with one model over another, or if they don’t see their (immediate or aspirational) selves in the image. Historical data! To know if you should go ahead with your top preference, you can consider looking at that data two-fold.

Retail Analytics

Firstly, a robust retail analytics portal like, say, Style Arcade allows you to filter by the model during any given period. You can choose as many filters as you want, but make sure they remain consistent as you flip through the analytics of different models.  

The model choice won't be the be-all and end-all for a product selling, it could have come down to a particular promotion, best-selling style, or colour. However, the key is to look for patterns. For example, in the units sold vs the returns rate.

In the example below, we have Cindy vs Naomi. Each with 6 styles at full price in the same period.

screenshot of Style Arcade app depicting statistics comparing models Cindy and Naomi

From 6 products it appears Cindy sells an average of 29 units per product, with an average return rate of around 2%. Meanwhile, Naomi sells an average of 55 units per product with an average 8% return rate. This generally tells us that the way both Cindy and Naomi wear the clothes, and how they’ve been shot is accurate to the size the customer predicted they would need. Naomi seems to be the winner here if you want to go by average units sold.

You can go deeper; if we look at the first 6 weeks of the product dropping on site, you might get a better read of how the audience responded to that model from the jump.

screenshot of Style Arcade app depicting statistics comparing models Cindy and Naomi over the 1st 6 week sales

Cindy sells an average of 4 units per product in the first 6 weeks, while Naomi sells an average of 6 products in the same time frame. Cindy had an average Sell-Through Rate (STR) of 20%, brought up by one dress that went gangbusters with a  STR of 78%. Naomi's average STR is 13%, and it’s all looking pretty consistent across the board for Naomi. I say, book Naomi again.

example of Style Arcade insights related to model product imagery
Style Arcade provides actionable insights relating to product imagery


Secondly, review-based data for brands with active online communities offer a world of insight. Free People allows their loyal community to go to town on what the garment fits, feels, looks, and wears like, compared to the way it’s shot on the model - but not without collecting a little zero-party data along the way. Touché FP, touché.

Case in point: Here we have a tank. This simple tank is an uncomplicated, common wardrobe staple.

web screenshot of Free People website showing model in black tank top and product details

It’s shot on the classic 20-something FP boho-beach-babe model in the studio, on location and even on a male model. It’s just a tank! Or is it? Not according to the reviews.

screenshot of Free People website showing 2 reviews for Kate Tank Top

These customers are so laser-focused on purchasing the perfect item, that while they’re willing to tell the seller the fit is off - they have voluntarily provided their body type, height, age and location, the size they bought, and the size they usually buy.  

Not only are other customers able to make more informed decisions, but Free People now knows the average age of someone buying this tank is much higher than 25. Free People also have the opportunity to reshoot the item based on the fact that their reviewers are saying the item runs small.

So, why don’t we have 50-year-old models on Free People? That brings us conveniently to our next question.

3. Have you considered your current audience?

Is your brand aspirational, or attainable? Both can be equally commercial, so it comes down to what your audience wants to see.

For every one of Kim Kardashian’s 325 million followers, according to the Skims selection of eCommerce models, everyone can be a Kim-bot. This kind of visual diversity is not only brilliant for the attainability of the world's most watched wardrobe - but it’s imperative for the brand to reach each style’s potential. For an audience that expansive, why wouldn’t you let them have their Skims and wear it too?

screenshot of Skims website product imagery showing sizing options and try before you buy
Try before you buy gets a whole new meaning, and is a whole lot faster

4. Considering User Generated Content

For many brands born on social media, there is a significant opportunity to leverage and encourage User-Generated Content (UGC) so that the brand story is continued across channels. Not to mention solving the ever-present comment on fashion influencer images: ‘where’s the top/shoes/pants from?’

Further to using UGC as a lead image, White Fox has tagged the Instagram personality in the product description, which works to give credit - but also provides a certain level of credibility (virality even) to the item.

screenshot of White Fox Boutique's website showing black oversized blazer on White Fox model and Instagram influencer

Realisation Par is pretty up there as a good example of a brand that merges its user-generated content, campaign, and eCommerce images to execute a seamless omnichannel experience.

Given there is no bricks-and-mortar in their equation, the onsite experience must heavily reflect their social media. Chances are (if you’re into floral silk dresses and summer in LA) you’ve seen the eComm imagery well before you’ve even reached the website.

screenshot of Réalisation Par's website showing The Stephanie dress in multiple polaroid images and product details

Additionally, by giving their eCommerce models the socially desirable ‘Dreamgirl’ prefix and referring to them by name, they’re instantly humanised to the degree that they are ‘stalkable’ on social media.

For social brands, eCommerce models (with a following) present an authentic opportunity to post themselves wearing the brand out and about or even better, on-set. Hello, hype. That’s when eCommerce imagery truly goes meta.  

Shooting your eCommerce models

1. Headless, partial or full body?

Not to be confused with headless commerce. According to Vogue Business, online retailers have found that straight-forward, front-facing model images, paired alongside an average of five detail shots, encourage purchases.

Seems straightforward, right? Only data will tell. Or Twitter users will.

tweet from @nyambura "Thanks Zara. This is exactly the post I wanna see for the coat" with image showing crouching woman with tan Zara coat over her body

With so many options to purchase online at any given moment, everything from the framing, how much of the face you show, the lighting and the styling can make or break the conversion rate on that item.

Although some e-commerce platforms follow the same guidelines for men’s and women’s, storytelling through product imagery is more common for womenswear, while men are more likely to shop using the left brain. As in, to get in and get out, even online. Not too much emotion is involved for someone on a mission for a very specific clothing item. Why do you think Reformation has these?

Men are also said to be more brand loyal, and less likely to experiment with new products which pretty much makes a well-styled model pretty useless at hooking them in. For example, according to Vogue business, Revolve doesn’t show any model’s faces on the men’s section of its site, as previous tests showed male customers were more drawn to images that focused on the garment itself.

2. What about video?

Both the main pro and con of eCommerce video: you cannot hide the fit in a video (serial bulldog clip stylists, beware!). This practice is likely to become even more common, with 73 percent of US adults saying they are more likely to purchase something after watching a video of it, according to the video creation service Animoto.

Videos under 30 seconds are said to perform best, but just make sure they’re rendered for web, and your site can handle the volume. No one likes to press play on a video that doesn’t load.

It’s vital as one of the definitive ways to get your product into bags, because when it comes to product imagery, the more you can mimic the experience of shopping offline and seeing a product up close, the better.

Final word: supporting your content

POV: You, the eCommerce manager, jumping on other sites to see what they’ve got going on. The eComm video load speed is smooth and the studio is well-lit. A+.

The eComm imagery is an obscure campaign shot. You have no idea how the fabric moves, or what the garment looks like from all angles.

There are so many ways you can influence your customer’s final decision with content supporting the imagery, including:

1. Product Descriptions

The bane of a brand’s existence, (ask a copywriter who knows) is product descriptions. However, there’s nothing more imperative in getting the customer across the line than the fabric and fit in the description.

Aside from any romance copy, ensure you’ve detailed at least the below:

  • the fabric composition,
  • the desired fit (boxy, relaxed, slim, cropped), and
  • whether or not the item runs true to size.

We know styling can get a little wild on set with the clipping of samples that are too big on your eCommerce model… but how does this translate for the customer? Is your model really wearing a size 8? Or a pinned 10? How do you plan on describing this fit to the customer?

2. Pre-empting customer questions

Try and pick up on things that might have been missed in the images, as they could lead to more questions than answers.

Is it a little bit sheer? Is it bra-friendly? Is anything detachable?

Did you shoot the sample shot, and now you’ve got a different product sitting on the rack in front of you?

These are questions you can’t afford to skip over.

3. Marketplace shoots

If you’re one of the thousands of brands operating on other stockists and marketplace platforms, do you know how you rate against the competition? Do your products on Farfetch look like they’ve been boxed up for a week (because they have). Do you actually just look like a knock-off brand of your own brand on  Amazon?

While you can’t choose their models, the best practice is to include styling advice or how-to-wear points for eCommerce shoots you can’t be around for. Like, if the bow should be tied at the front, whether it’s able to be worn multiple ways or what other products from the collection it should be styled with.

Interested in learning more about how retail analytics can drive ROI by making sure you're selecting the right product model? Get in touch with one of our team today.  

Subscribe to Style Arcade
Arrow pointing north-east