How has the Coronavirus changed product imagery?

Just how much has model usage dropped—and have mannequins picked up the slack?

And how much has the impact varied by product category, region, and size of the business involved?

In this blog post, we’re going to share the changes we’ve seen in the images we edit: the drop-off, recovery, and type of imagery we’ve been retouching.

Using Pixelz retouched images as a proxy for the industry as a whole isn’t perfect, but we have enough volume that we believe our insights are an interesting indicator.

At Pixelz, we retouch many millions of e-commerce images every year for brands, retailers, and commercial photo studios around the world. This July, we edited our 39 millionth image.

And because we have developed an AI-assisted digital lean production system that breaks retouching down into component steps, we have extremely detailed information on the images we process and related Photoshop activity. Literally terabytes of data.

Our hope is to give hope—yes, the industry has been hit hard. But the recovery is well on its way, and there are practices (like using mannequins or vendor images) that can mitigate challenges.

UPDATE: This blog post was originally published on August 10th, 2020. Pixelz has since more than fully recovered all lost image volume and is in fact setting records—but how much is due to new business versus recovered business would require deeper data analysis. The insights in this blog post are still valid, but should be understood as reflecting a moment in time during a rapidly changing environment.

Let’s get to it.

Table of Contents


In March of 2020, COVID-19 blew up across Europe and North America and e-commerce image production went off a cliff.

While some companies saw it coming and took precautions, like frontloading model shoots in February and early March, or sending equipment home and planning for remote shooting, the industry as a whole took a huge—and sudden—hit.

Total Image Volume

At Pixelz, we saw the drop happen dramatically day by day as lockdowns went into effect. Our image volume dropped by 27% on March 17th, was down to -49% by March 19th, and stayed at about -50% the following week.

There are many millions of images represented in this chart, so any noisy elements should be washed away.

weekly product image retouching volume during COVID

Weekly total e-commerce image volume

Instant analysis

  • 50% drop in mid to late March
  • Bottomed out in ~4 weeks
  • 25% recovery to-date

Narrative analysis

Okay, we see the cliff and now (yay!) the climb. But we need to drill down.

Where did the drop-off and recovery come from?

Was it spread across customers, or did some customers go to zero while others were unaffected?

The answer is that almost everyone was affected to some degree—and that fortunately, almost everyone has also recovered to some degree.

Total Active Customers

Our 50% total drop in image volume came from a combination of a 30% reduction in active customers and a reduction in volume from those active customers.

An “active customer” is someone who uploaded images for processing.

This is a binary metric: either you’re active or you’re not.

Pixelz has 30,000+ customers in 100+ countries around the world.

Number of active customers per week during COVID period

Active customer count (uploaded any images during the week)

Instant analysis:

  • 30% drop in March-April
  • 21% recovery by week 22 (May 18-24th)

Narrative Analysis

The good news is the “active” drop was much less than the image volume drop, and bounced back relatively quickly.

This seems to indicate that for a week or two, people were scrambling hard to find ways to produce imagery. And some were scrambling so much they weren’t able to produce at all—but as soon as they could, they did.

So COVID hasn’t been a permanent shutdown for most.

But who dropped out?

Whose business hasn’t come back out of lockdown?

Who Went to Zero?

We can parse this based on our clients’ tier of service.

At Pixelz we offer three different plans, serving clients with different scale and needs.

  • Solo - For solopreneurs with low volume looking for essential image edits like background removal, cropping, alignment, etc.
  • Professional - For busy retailers who have higher volume, a need for speed, and more sophisticated retouching requirements.
  • Enterprise - High volume brands and retailers who need a custom experience at scale.

(compare plans)

Below we look at active customer count by plan type to see who suffered the most dramatic event of all: going to zero.

active retouching service customers per week by plan type

Active customers by plan type: the lowest image volume tier took the biggest drop and has not fully recovered.

Instant analysis:

  • Solo dropped by 40%, has recovered to -15%
  • Professional dropped by 35%, has fully recovered and even grown
  • Enterprise was unchanged throughout

Narrative analysis

In broad strokes, the smallest businesses stopped producing imagery for a time. Perhaps product stopped arriving for a small business, or a boutique that closed its doors in real life closed them online too, or there was no time for a side hustle during crisis.

But for larger SMBs—those on our Professional tier—after an initial scramble, the critical need for e-comm imagery kept images being produced. And the re-thinking of workflows and need for efficiencies may have actually led more brands and retailers to look for retouching services, resulting in growth for the segment.

And finally, enterprise level brands and retailers may have experienced drops in image volume, but all of them stayed active and kept producing imagery for e-commerce. No surprise that the only channel that could stay open was treated as business critical, and that clients stayed with a service that had no operational impact from COVID-19 (read about ourrisk mitigation architecture).

But let’s be clear—although no Enterprise client went to zero, they did experience contractions too.

Let’s take a look at that.

Image Volume by Business Size

Again, we'll use Pixelz plan type as a proxy for business size.

line chart of weekly retouching image volume change over time by Pixelz plan type

Images uploaded by Pixelz plan type. Bigger businesses are recovering faster than the smallest.

Instant analysis:

  • Enterprise dropped by 50%, recovered to -20%
  • Professional dropped by 60%, recovered to -20%
  • Solo dropped by 40%, recovered to -30%

Narrative analysis

Those with the highest volume saw the biggest drop-off. It’s impossible to replicate the productivity of a high volume professional studio when shooting remote.

But there are efficiencies to be found, and studios have re-opened in different regions—or brands have built studios in less impacted locales, be that Idaho or Hong Kong. As that’s happened, the bigger guys have started to recover.

Not so the little guys. As seen earlier in our “who went to zero” chart, the smallest businesses were heavily impacted. There’s not much of a positive trend line to suggest they’ll be coming back soon.

Image Volume by Region

Geopolitically, nations (and regions within nations) have responded quite differently to COVID-19. Lockdowns’ length and severity varied widely.

How did that play out in terms of photo studio production?

line chart of retouching service volume during covid by region: north america, scandinavia, and EU

Note: Pixelz was founded and is headquartered in Denmark, so we have a deep Nordic / Scandinavian customer base.

Instant analysis

  • North America dropped 60%, recovered to -25%
  • Scandinavia dropped 50%, recovered to -15%
  • The rest of the EU dropped 65%, recovered to -25%

Narrative analysis

North America had the sharpest drop, likely driven by studio concentrations in New York City and Los Angeles. Both cities went from fully open to locked down almost overnight.

Scandinavia saw a more gradual drop and a more complete recovery. The recovery might be even more complete, but we’ve begun to run into the Nordic slow season, where it’s common for citizens to take at least 3 weeks of summer vacation (yes, I’m jealous).

The rest of the EU isn’t too dissimilar from North America in scale of drop and recovery, which is somewhat surprising. Because the virus struck earlier than in North America, I expected we might see earlier and more complete signs of recovery. It’s possible we are again running into a summer slowdown.

The blending of regions like Germany, Benelux, and the UK, which have had quite different impacts from the virus, keeps us from trying to draw too many conclusions. But I will say when looked at individually, most EU country lines were not too different—hence our decision to blend them.

Model vs Mannequin, Global

Pixelz is an e-commerce retouching service. Our biggest segment is the fashion industry, especially apparel and footwear.

As mentioned earlier, because Pixelz has developed a proprietary production system that blends specialist editors and automated processes, we know exactly the type of retouching performed on each image.

The data presented below is based on the retouching steps an image went through. It’s not straight classification: so for example, an image that has a model in it but for which model retouching was not requested will not be counted, and the same for mannequin usage.

The most common mannequin specific retouching steps are smoothing of shape and the creation of the Ghost Mannequin effect.

line chart of model retouch vs mannequin retouch over time during COVID

Model retouching vs mannequin shape retouching and/or ghost mannequin

Instant analysis

  • Mannequin shape retouching dropped by 70%, recovered to -25%
  • Model retouching dropped by 75%, recovered to -45%

Narrative analysis

You probably expected Model usage to be the most impacted part of product photography. You would be right, and we’ll look at that in more detail next.

Mannequin usage has recovered, but it would be inaccurate to say there’s been a surge in mannequin usage, or that it’s picked up the model slack. Mannequins still haven’t fully recovered, but their recovery is on par with the industry’s overall recovery (total image volume, by region, etc.).

It’s speculative, but to me that indicates that even though mannequins are a viable option for fashion photography at scale in the remote studio era, remote studios just aren’t as efficient.

Model Usage by Region

As we said last graph, you would probably expect lockdowns and social distancing to have dramatically impacted model usage.

You would be right.

Model usage is sensitive to a number of factors: they’re expensive, take time to book, may involve travel, and require a lot of close personal contact in the form of styling and hair and makeup.

In the United States, models are also disproportionately located in New York City and Los Angeles, both cities that had strict and prolonged lockdowns. Photo studios are often located in those cities in large part for access to talent.

So you can’t relocate model shoots to Iowa quite the same way you can mannequin photography.

Let’s look at model usage by region.

line chart of model retouch by region over time during COVID

North American (USA & Canada) model usage crashed fastest and farthest, but has since climbed steadily

Instant analysis:

  • North America dropped by 95%, recovered to -30%
  • EU minus Scandinavia dropped by 90%, recovered to -30%
  • Scandinavia dropped by 80%, recovered to -20% (then dropped again as summer vacation hit)

Narrative analysis

North American model usage dropped both deeply and extremely sharply, bottoming out in 3 weeks, but has been slowly and steadily recovering. At time of publication, we’re on a streak of 5 consecutive weeks of growth, some of which were quite significant.

It is worth noting that Canada makes a real impact on the North American trend lines: Canadian studios went to near zero and have stayed there, as the country has adopted a much stricter approach than the USA.

Scandinavian model usage dropped sharply, but it also rebounded quickly. They were back up to roughly 75% of prior volume within 5 weeks. The recent volume drop can be attributed to the aforementioned extended Nordic summer holidays.

The rest of the EU has had a bumpier ride, which makes sense considering the variety it encompasses. After taking 9 weeks to bottom out, the EU’s been on a mostly positive trend except for two weeks of notable reversal in the middle of June.

Finally, I’d like to point out that each region saw a deeper drop than Pixelz ever did as a whole. Despite regional drops of 80-95%, our global model retouch low point was around 75%. The line is smoothed by when regions hit bottom.

The really positive takeaway is that each region has seen significant recovery and is on a positive trend (if you disregard the Danish summers).

Alright, that gives us some insight into regional impacts. But who was most impacted by the loss of models?

And who’s shooting on-model now?

Model Usage by Business Size

As we saw last graph, lockdowns and social distancing dramatically impacted model usage.

Let’s look at how that impact was distributed by size of the business.

If you’ll recall, we have three tiers of service which roughly map to business size. We’ll leave the smallest tier out of this, since they shoot much less on-model.

comparison of model retouch service usage by business size during COVID

Weekly images processed with model retouching: 85% drop, approx. half recovered

Instant analysis:

  • 85% drop in model usage by Enterprise, recovered to -55%
  • 80% drop in model usage by Professional, recovered to -40%

Narrative analysis

Person to person spread of COVID makes shooting with models on a professional set extremely challenging. Styling, hair and makeup, these are all things that can’t be done socially distanced.

Anecdotally, budgets are tight and there’s also been more audience acceptance of influencer imagery and other less traditionally produced content.

But we have seen some recovery, with 10% greater recovery at the Professional tier than at Enterprise. This suggests that it’s easier to shoot on-model at smaller scales.

Why might that be?

There's quite a few possibilities.

One, we've touched on previously—even if an Enterprise operation were to go back to shooting fully on-model in their normal studio space, they're just not going to be as efficient as they were before due to COVID distancing. Less space, less staff, more time for cleaning, etc. In contrast, a smaller operation was probably not as optimized to begin with.

Perhaps it's easier for a smaller business to adjust their style guide and standards on the fly than it is for a giant corporation with a more established aesthetic.

It may be that risk assessment and liability differs at scale, as does just how business critical it is to get those on-model images up.

Most likely it's some combination of the above and a host of other factors.

In Summary

E-commerce image production took a beating in March, but production has slowly been climbing back to its previous heights.

Pixelz has seen half of our image volume loss recovered, in a steady trend since early April.

On-Model photography in particular suffered a steep decline, but it too is steadily recovering across regions and business sizes.

Enterprise level companies are recovering faster than scrappy boutique startups, but mid-tier businesses dropped farthest and are recovering fastest of all. Those SMBs are seeking out efficiencies, including retouching services like Pixelz, and are finding ways to shoot on-model again.

The smallest businesses seem the most at-risk of an incomplete recovery. On a percentage basis, more Solo clients were lost than any other tier, fewer came back, and their image volume has seen the least recovery.

In the event of a second wave, or the return of quarantine restrictions as the USA is seeing in states like California, it seems unlikely the industry will see such a steep drop again. Studio teams now have experience handling remote studio setups and risk mitigation plans are in place.

E-commerce continues to be a critical channel, and is in fact more important now than ever before as brick and mortar faces unprecedented operating restrictions.

So the sky remains cloudy.

But it’s brightening.