WordPress multisite adds operational complexity to image optimization. One network may contain marketing sites, regional sites, help centers, product microsites, franchise pages, and internal portals. A bulk WebP run across the whole network can affect many teams at once.
Before converting media across multisite, plan ownership, batching, backups, permissions, cache behavior, and per-site quality standards.
The main risk in this topic is oversimplification. "Run a bulk optimizer across the network" is not a plan. A useful multisite plan states which subsite is being changed, who owns it, which license and quota apply, how delivery is verified, and how that subsite is rolled back without disturbing the rest of the network.
Inventory Sites Before Files#
Start with a site inventory, not a media inventory. Each site may have different traffic, templates, plugins, image roles, and business owners.
Record:
- site name and URL
- owner or approver
- traffic priority
- theme or template family
- ecommerce or product use
- known media-heavy pages
- cache and CDN behavior
- special compliance or privacy needs
This helps decide rollout order. A low-traffic brochure site and a high-revenue store should not receive the same risk treatment.
For GetWebP specifically, anchor the inventory to the multisite support documentation: each site has its own settings, queue, license key, and quota counter. Even if the plugin is network-activated, the operational record should still be per site.
Confirm Media Storage Structure#
Multisite media paths and attachment records can be easy to confuse. Files may be separated by site, while shared themes and plugins may generate similar names. Review where each site's uploads live before running a tool.
The WordPress Media Library documentation is useful for the editor-facing model, but multisite operators also need the network-level file layout and permissions.
Never run a conversion against a shared path until you know which site owns each file.
Add a storage row to the rollout sheet:
Site URL: https://example.com/uk/
Upload path checked: yes
Attachment ownership checked: yes
Shared theme image folder excluded: yes
Offload/CDN path checked: yes
Backup scope: site uploads + site tables
That level of detail prevents a batch job from touching shared assets that are used by another subsite or theme.
Define Network-Wide Defaults and Site Exceptions#
A network can have default image rules, but exceptions should be expected. For example:
- marketing sites use WebP for hero and blog images
- ecommerce sites preserve higher-quality product zoom files
- documentation sites treat screenshots carefully
- brand sites keep logo files lossless or SVG
- internal sites may need stricter privacy handling
Write the defaults, then document exceptions per site. This reduces surprise when a site owner reviews outputs.
GetWebP's batch conversion documentation lists the named presets and the fact that generated WebP/AVIF files are sibling variants, not replacements for originals. Use that as the default policy language, then make exceptions explicit:
| Site type | Default | Exception to record |
|---|---|---|
| Marketing microsite | Balanced preset, WebP and AVIF siblings | campaign hero approved by brand owner before launch |
| Ecommerce store | Higher visual review for product zooms | do not approve if color or labels shift |
| Documentation site | conservative handling for screenshots | reject output if UI text becomes fuzzy |
| Brand center | SVG or lossless source retained | no lossy conversion for logo masters |
| Internal portal | verify privacy and access rules first | do not include restricted folders in screenshots or reports |
Plan Backups by Site#
Backups must match the recovery need. If a conversion fails on one site, the team should know whether it can restore only that site's uploads and database records or whether a wider network restore is required.
The WordPress documentation on backups is a useful reminder that both files and database matter. In multisite, the restore process may involve site-specific tables, shared network data, and uploads folders.
Test the restore path before a network-wide job.
Batch by Risk, Not Just Count#
Do not batch multisite work only by number of files. Batch by business risk and template similarity.
A practical sequence:
Batch 1: staging copy of one low-risk site
Batch 2: two similar low-traffic marketing sites
Batch 3: high-traffic blog after approvals
Batch 4: ecommerce sites with product-image review
Batch 5: archive sites and cleanup candidates
This lets the team learn from easier sites before touching sensitive catalogs or campaign pages.
If WP-CLI is available, dry-run each subsite before enqueueing real work. The WP-CLI command reference documents wp getwebp convert, wp getwebp status, and the standard multisite --url targeting pattern:
wp getwebp status --url=https://example.com/uk/
wp getwebp convert --url=https://example.com/uk/ --dry-run
wp getwebp convert --url=https://example.com/uk/ --format=webp
Do not replace this with a blind network loop until the first sites have passed visual QA, delivery checks, cache purges, and rollback rehearsal.
Check User Roles and Editorial Workflow#
Multisite often has distributed editors. A network admin may control plugins, but local editors manage media and content. If optimization changes what editors see or upload, communication matters.
Check:
- who can upload media
- who approves visual quality
- who can replace files
- whether local admins can bypass network rules
- how exceptions are requested
- whether training or documentation is needed
Technical success still has to leave editors with a workflow they can maintain.
Verify Cache Layers Per Site#
Some sites in the network may use different cache settings or CDN rules. A WebP rollout that works on one subsite may fail on another because of page cache, path rules, or edge behavior.
For each batch, test:
- public page HTML
- selected image candidate
- response
Content-Type - CDN cache status
- fallback behavior
- cache purge process
Do not assume network-wide plugin activation means network-wide delivery consistency.
GetWebP's frontend delivery documentation matters here because multisite networks often mix themes and cache rules. Verify whether the subsite uses direct src rewrite, picture insertion, output buffering, server rewrite, or a CDN path that needs its own purge.
Keep a Network Rollout Record#
Use a record that tracks both site and file status:
Site: /uk/
Batch: marketing pages
Activation: per-site
License/quota: checked
Dry run: 90 candidates
Files converted: 84
Skipped: 6 screenshots
Preset: Balanced
Format: WebP only for first pass
Approver: regional marketing
Visual QA: passed hero, product card, and article template
Delivery QA: picture fallback checked on public pages
Cache purge: complete
Rollback path: site uploads restore
Status: approved
This helps network teams answer questions without rerunning audits.
WordPress multisite image optimization is an operations project as much as a compression project. Plan by site, preserve ownership, test backups, batch by risk, and verify delivery for each site before expanding the WebP rollout.

Jack
GetWebP EditorJack writes GetWebP guides about local-first image conversion, WebP workflows, browser compatibility, and practical performance checks for teams that publish images on the web.