How to Edit E-Commerce Product Photos for Higher Conversions

The process of product image background removal isolates the item from its background and adds an acceptable background as per e-commerce site (or the brand). In photo editing for online stores, removing the background is the very first phase. All subsequent processes such as creating shadows, correcting colors, and composite can only be done after product image background removal.

Product image optimization requirements set by platforms cannot be ignored. For example, Amazon and Walmart require all primary listing photos to be shot against a solid white background (RGB 255, 255, 255).

Secondary images on Shopify and Etsy may use lifestyle or contextually relevant background images; however, according to numerous studies, and controlled AB testing, using a white-background primary image always performs better than alternative background types.

There are three ways to remove the background depending upon the complexity of the product being photographed.

• Clip Path/ Pen Tool: This method is used for hard-edge items: shoes, electronics, glassware and packaging. Clip Paths provide sub-pixel accuracy and are required if the edited asset needs to go to print or be displayed at high magnification or need to be sent back to an agency.
• AI-based Removal: Adobe Firefly, Remove.bg, Clipping Magic provides quick and effective results for items with distinct silhouettes. However, these programs do require manual quality control on difficult-to-remove areas of the product: hair, fringe, mesh, and transparent materials.
• Color Range/Magic Wand: This method only works well if there is one consistent tone for the entire background. Both tools fail on products with shadows, gradations of tones, and any bleed from the product onto the background.

Following AI-removal of the background, open up the Select & Mask Workspace (Photoshop’s Refine Edge) to refine any difficult edges from the previous steps. Halos around objects that were once part of the background, and any remaining jagged cut outs or fringe are typically identified during this step. Any of the above-mentioned errors are viewed at full zoom and would result in a rejected image due to platform standards.

With the background removed accurately now begins the process of adjusting colors which becomes the second highest priority and the largest driver behind a 54% return rate making color correction the most commercially impactful tool in the workflow.

Color correction ensures that the colors of images are accurate when compared to the physical products they represent. The color correction process involves adjusting the Hue, Saturation, Luminance, and White Balance to ensure that the colors shown in the images accurately reflect those seen in person when illuminated with neutral light.

A correctly calibrated and hardware-profiled monitor (sRGB or AdobeRGB) is necessary for this process along with a carefully managed step-by-step process. It is not acceptable to simply make a global Hue/Saturation adjustment.

Steps one through four:

First Step

Set Neutral Reference Point

Use the Level’s Eyedropper Tool to capture a color from a known neutral object such as:

• White product label.
• A gray Card taken during the photo shoot.
• A white Background Reference.

Second Step

Apply Curves:

Independently adjust the highlight, midtone, and shadow areas using curves to increase the number of tonal levels within image while avoiding adding additional color cast to the area being adjusted.

Third Step

Use Selective Color (CMYK Sliders):

Individually adjust the color of objects within your image (without changing the surrounding tone) using CMYK sliders. For example, in the world of fashion and cosmetics a two-point adjustment in Cyan will turn a Cobalt into a Periwinkle and result in a color-mismatch which results in returns.

Fourth Step

Match Colors Between Variants:

Sample the color information from an approved Hero Image(s) and apply that color information to each of the Variant Images. This eliminates the need for individual correction of each variant image and allows to maintain consistent visuals throughout the entire color-swatches row on the product listing.

Do Not Over Saturate Products

While over saturation may be perceived as visually appealing, it often creates a perception that product image has been manipulated in flat product categories, and in some of the most trusted industries (jewelry, cosmetics, food etc.) can create negative reviews due to the fact that consumers trust what they see and want to buy based upon an accurate representation of what the product actually looks like.

Ghost Mannequin photo editing is commonly referred to as “invisible mannequin” or “hollow man”. This method involves applying a post-production composite effect to remove the mannequin/model from the apparel photograph.

After removal, the resulting image depicts the garment as if it were standing alone in 3-dimensional shape. Thus, the viewer can see how well the garment fits, hangs, has a defined collar area and hem line behavior all this without the distraction of the mannequins/forms. Additionally, there is no need for each garment to have a live model photographed.

To accomplish this process, we require a minimum of 2 photos per garment; one exterior photo of the garment on a mannequin and another interior photo of either the neck area or a lining detail of the same garment. Once these photos have been layered in Adobe Photoshop, masks are applied to reveal the neck and hem areas of the outer garment.

Using frequency separation retouching techniques, inconsistencies (tonal steps-changes/edge artifacts) are eliminated that may be present at the composite junction caused by our previous actions.

This process is now common practice among most DTC (Direct-To-Consumer), Private Label Fashion Sellers, and Retailers that sell high SKUs (stock keeping units).

We here at Hitech Digital process approximately 2,000 to 3,500 apparel photographs daily for U.S. based retailers using a Rule-Based Workflow for processing Ghost Mannequins. Our documented error rate is less than .03% and our Standard Level of Service (SLA) is a 24-hour turnaround time.

For further information about this process, please review the case study.

Products without shadows appear as if they are floating above the product image. This creates distrust in consumers to make purchasing decisions. Adding a drop or cast shadow to your images creates a visual connection to the product and provides consumers with the perception of an actual 3D product.

There are three different ways to create shadows for different uses:

• Drop Shadows (default for all product categories): A soft, faint shadow located under the product.
• Cast Shadows (standard for footwear, furniture and products shot at angles): A long, directional shadow that matches an identified non-directive light source.
• Reflection Shadows (used for glassware, ceramic products, polished metals and electronic devices with reflective surface): A vertical reflection of the product image that has been distorted, blurred and faded.

The only way to provide a realistic look to these three types of shadows is to ensure that the direction of each type of shadow corresponds with the direction of the lighting used in the photo shoot. Irregularity within catalog is a quality control issue; irregularities undermine the overall credibility of a catalog – not only individual photos.

A clipping path is an image of a line defining a single, enclosed shape (vector), created in photoshop. The clipping path defines what portion of the image lies within the area defined by the path. All pixels lying outside of the clipping path are removed from the exported file; therefore, it is a technically accurate method for delivering files to ad agency, print vendors, or e-commerce sites that require products to have a clear border around them and no residual “background” artifacts.

The process of using multiple clipping paths allows you to assign a name to each separate clipping path for individual colors in your product. If you were working with a shoe with three different colors (white midsole, black bottom sole, blue top), you would create three separate clipping paths allowing you to isolate/swap/correct each of the colors individually without changing any other color.

This is how most online retailers manage their color-variant catalogues without having to reshoot each variation. It’s the only way to do so economically at scale.

A good lighting and exposure adjustment is about maintaining your picture’s dynamic range. In other words, it involves keeping both the dark areas of your picture (shadows) and the lightest areas (highlights) from being clipped or lost. When you lose texture in a product image due to having blown out high lights, the consumer may miss important product characteristics when evaluating the product.

On the flip side, if you block up the shadows so that they do not appear as dark as they really are, you hide product features the customer will be looking at when deciding whether to buy the product.

Workflow:

• Adjust Black and White points in the levels tool, prior to adjusting any other tool, to expand the overall dynamic range of image.
• Use curves to adjust exposure by independently controlling each part of images’ tone scale. The subtle “S” curve provides a good amount of contrast while providing better output than using the Brightness / Contrast sliders globally.
• Dodge and Burn on a 10-15% opacity overlay layer to refine specific parts of image such as labeling text, hardware details and/or fabric weave textures.
• Use Camera Raw to recover blown-out highlights. RAW files provide 2-3 stops of recovery room whereas JPEGS have no recovery room.

Editing: Removing Defects without misrepresenting the Product Editing (retouching) removes handling marks, fingerprints, dust, scratches and transit damage that occur from factory to photo studio. In no way can it be used to alter the products manufactured design, texture or structural elements. This is the greatest commercial constraint within the entire editing process. Select your tool based on the defect you are trying to correct:

• Spot healing brush – random surface markings on flat surfaces with a constant finish.
• Clone stamp – larger structural areas where there will need to be an exact replication of surrounding textures.
• Healing brush/patch tool – fabric wrinkles & packing creases. Preserve natural folds/drapes and only remove creases that were created due to shipping/handling.

Over-retouching – smoothing out intentionally textured materials, concealing structural seams/boundaries, lightening material grades — causes over half of returns (54%) as noted in section one. Often times the difference between the image being presented and the actual Product received is not a photography issue but rather an editorial decision which crossed into Product misrepresentation.

HitechDigital completed and retouched 50,000 images of jewelry for a pa brand improving brightness/contrast/color correction and removing all surface imperfections resulting in tangible improvements to their customer engagement/conversion metrics. View our jewelry editing case study.

High-pass sharpening should be used as the first choice for a non-destructive approach to sharpen an image. It can be achieved using the “high pass” filter with an empty duplicate layer in the top layer of stack. Set this top layer to either hard light or overlay, and you will have the ability to adjust the level of sharpening by adjusting the layer opacity. The high pass sharpening tool allows you to target high-contrast areas while avoiding the introduction of halos on flat, low-contrast surfaces.

You can improve clarity and sharpening in a product image but they are targeting two separate issues.

Clarity (midtone contrast) helps increase the visual depth perception of a surface. Sharpening helps define sharpness of edges. However, if you add too much clarity to a high-contrast area, it may produce halos similar to what excessive sharpening does. If there is no need to use clarity in a particular surface area due to damage to the detail, try to isolate and make local curves changes to the particular range(s) of tones.

Keystoning (the product looks wider on top than it does on the bottom) can occur from perspective distortion. This happens because of how far away the camera sensor is from the subject (product), creating a non-parallel relationship between the two planes, thereby making the product look like it is widening down.

Keystoning will always happen if a macro set up is used and when the camera is placed directly above or below your products center line. Lens Correction Filter in Photoshop allows for either automatic corrections by utilizing built-in lens profile information, or manual adjustments through a grid-based system.

The Vanishing Point filter within Photoshop provides greater flexibility in terms of placing the convergence points for products with strong geometric lines (packaging, electronic components, furniture). GIMP uses the Perspective Tool and the Distort › Lens Distortion tool to accomplish similar tasks. There is no “correct” amount of correction regarding product geometry.

As long as all parallel edges look parallel after you have applied the corrections, you do not need to make additional corrections. Applying additional corrections will create a false flat perspective that many customers may view as having been manipulated photographically as opposed to being photographed.

Alignment & Cropping help remove unnecessary dead space from your image subjects; alignment places your image subject at the same position inside a predetermined boundary on each SKU in your catalog.

How to do it

Determine the aspect ratio you want to use for all images before you begin the cropping process. The majority of e-commerce platforms (e.g., Amazon) have a 1:1 aspect ratio. Editorial/lifestyle images can use a variety of ratios such as 4:3 or 16:9. Once you determine this aspect ratio, perform a bulk crop operation for all images to ensure the ratio is locked throughout the entire catalog.

Align your product intentionally

Position your product intentionally based upon the type of shot you are creating:

• Hero shots → Center your subject. This allows the viewer’s eye to focus directly on the subject without being distracted.
• Lifestyle shots → Use the Rule of Thirds when positioning your subject. This will provide enough negative space for overlayed text, while still keeping the product visually prominent.
• Tall products (e.g., Bottles, Apparel on Hangers) → Align the bottom of the product with an internal guide so that variations in height are perceived as differences in product rather than poor photography.

Perform a quick check

A simple method to verify that all images have been aligned correctly is to quickly toggle through multiple images. If there are any inconsistencies in the size or position of one image compared to another, it becomes immediately apparent.

Tools to help with these tasks

• Lightroom → Sync Batch Crops
• Capture One → Precision Guides for Aligning Images
• Canva → Grid Template For Rapid Edits
• Bulk Image Crop → High Volume Catalogue SKU Processing

Why it matters

Inconsistent frames create “visual jumping” for customers who view multiple images of your product offerings in a row – essentially creating subtle friction that can kill confidence prior to reading your content. Uniformly cropped and aligned images signify professional quality and build subconscious trust that ultimately leads to increased conversions. Proper cropping also enhances mobile usability, strengthens catalog consistency, and keeps the shopper’s attention on the product versus distractions created by poorly framed images. Additionally, well-aligned product grids enhance the overall look and feel of your storefront and contribute to additional engagement opportunities and higher conversion rates.