A 100/100 screenshot is reassuring. So is a checkout that still works. When you hire a WordPress speed optimization expert, ask for evidence of both: faster pages and a website that still does its job.
The short answer: Choose a specialist who investigates the bottlenecks before prescribing fixes, agrees on the pages and measurable targets in writing, and shows before-and-after evidence under comparable conditions. Ask how they protect your site, test important customer journeys, document changes, and handle problems afterwards. Relevant experience and a clear explanation of limitations tell you more than an unexplained perfect score.
Whether you are comparing my WordPress speed optimization services or somebody else’s, apply the same standard. You are buying an improvement you can verify, understand, and keep using.
In this article
How to choose a WordPress speed optimization expert: the short checklist
Why a 100/100 promise needs context
Ask how they diagnose before they prescribe
Look for experience with websites like yours
Ask for before-and-after evidence you can compare
Check how they protect functionality while improving speed
Get scope, ownership, and support in writing
How to choose a WordPress speed optimization expert: the short checklist
Start with these seven criteria. A provider need not produce a novel before being hired, but the proposal should make the work understandable.
| Criterion | What to ask | Evidence to request |
|---|---|---|
| Diagnosis | What appears slow, and what needs investigation? | A baseline and a reasoned starting point. |
| Relevant experience | Have you worked with similar functionality? | A relevant example with the work explained. |
| Measurement | Which pages, devices, and conditions count? | A written test plan and comparable reports. |
| Implementation safety | How will changes be deployed and reversed? | A backup, rollback, and deployment plan. |
| Functionality | Which customer journeys will be checked? | An agreed list and recorded test outcomes. |
| Handover | What will I receive and retain? | Change log, settings, exclusions, and next steps. |
| Support | What happens if an optimization causes trouble? | Written coverage, duration, and contact route. |
A clear answer can be modest. “This fee covers one landing-page template and its enquiry form” is useful. “Complete optimization” needs considerably more explanation.
Ask for evidence at each stage. This is an evaluation framework, not a provider rating.
Why a 100/100 promise needs context
A Lighthouse performance score of 100 can be genuine and useful. It describes a page tested under particular conditions; it does not certify every visit or prove that checkout works. Ask what was measured before deciding what the number is worth.
PageSpeed Insights combines two kinds of evidence. Its lab test uses Lighthouse in a simulated environment. Its field data comes from real Chrome users where sufficient data exists. A strong lab result does not automatically establish a good real-user experience. Google explains the lab and field distinction.
Core Web Vitals cover loading, responsiveness, and visual stability: Largest Contentful Paint (LCP), Interaction to Next Paint (INP), and Cumulative Layout Shift (CLS). Time to First Byte (TTFB) helps investigate initial response delays, but is not a Core Web Vital. Google’s Web Vitals guide separates these measures.
Likewise, lab Total Blocking Time (TBT) is useful diagnostic evidence; it is not a measurement of field INP. A report should not quietly exchange one for the other. See Google’s explanation of TBT.
Even lab results vary between runs. Lighthouse calculates a weighted score from individual metrics, so ask to see those metrics and the testing conditions. Lighthouse’s scoring documentation explains why the headline number moves.
Useful commitment versus vague promise
Illustrative vague promise: “Guaranteed 100/100 and a lightning-fast website.”
Illustrative useful commitment: “We will agree targets after the baseline, test the listed templates on mobile and desktop under recorded conditions, and deliver the results alongside functional checks and unresolved limitations.”
The second still needs actual targets, acceptance criteria, exclusions, and remedies in the proposal. If a provider offers a score guarantee, read those details. A defined guarantee can be legitimate; an undefined one leaves too much to the imagination.
A homepage screenshot may be accurate while saying little about a product configurator or member account. A desktop result is not a mobile result. Neither gives you a promise of rankings, revenue, or conversions.
Ask how they diagnose before they prescribe
A WordPress speed optimization specialist should be able to explain the next investigation in language you understand. Ask whether the initial wait, visible content, or response to a tap is the main problem—and what evidence would distinguish the possible causes.
A slow server response may need investigation of hosting, application work, database queries, or an external service. A page that arrives promptly but appears slowly may need attention to images, styles, fonts, or JavaScript. Third-party widgets can add work of their own. These are investigation routes, not a shopping list.
For an initial view, Speed Analyzer, which I develop, brings together performance checks and reporting. A tool can help establish the baseline; interpreting the cause and deciding what to change still requires judgment.
Ask what a proposed hosting move or plugin replacement is expected to fix. “Your current setup is bad” is less useful than a finding you can inspect. My explanation of why WordPress can stay slow after installing a caching plugin covers the distinction.
Tool choice should follow the website’s needs. The FlyingPress vs WP Rocket comparison is useful background, but a preferred plugin is not, by itself, a diagnosis.
Look for experience with websites like yours
“WordPress” covers a remarkably broad collection of things that can go wrong. A simple brochure site and a multilingual store with live shipping quotes may share a dashboard while needing quite different work.
Ask about your page builder, important templates, custom features, and integrations. For WooCommerce, name product variations, cart updates, and checkout. For memberships, name login, account pages, and protected content. Public cached pages and logged-in or dynamic behaviour need separate consideration.
Request an example with an explanation: what was slow, what changed, what was tested, and what remained limited. A redacted report can be useful when client confidentiality prevents sharing access. There is no need to demand somebody else’s private dashboard.
A large portfolio or a long career can support confidence, but neither establishes fit on its own. Listen for how the provider connects previous work to your actual requirements. The most useful sentence may be “That part needs further investigation.”
Ask for before-and-after evidence you can compare
Before you hire a WordPress speed optimization expert, agree how success will be checked. Otherwise, you risk comparing different pages, different conditions, and an impressive amount of enthusiasm.
Request the same URLs before and after, with the tool and version, device profile, network settings, test location where available, and dates recorded. Keep mobile and desktop results separate. If the tool chooses its own location, record it and disclose differences instead of suggesting that it was controlled.
Document login state, browser-cache conditions, server or CDN cache state, and relevant configuration changes. A content delivery network (CDN) can change delivery conditions; “CDN enabled” alone does not establish that a request was served from cache.
Ask for repeated lab runs and a stated summary method. For example, three runs per consistent condition with the median reported is a practical starting point, not a universal rule. Keep the individual results so an unusually good or bad run remains visible. Do not combine different cache conditions into a single reassuring average.
Choose representative templates: perhaps a homepage, article or service page, and product page. Then add the journeys your business depends on. A loading report and a checkout test answer different questions; both may belong in the handover.
Field results need context too. PageSpeed’s real-user data represents a trailing 28-day period, so it can still include visits from before the work. Low-traffic pages may lack data, and origin-level results are not a measurement of one specific page. Label the coverage clearly rather than promising an immediate field-data pass. Google documents the collection window and availability limits.
A screenshot is one piece of evidence. Request the surrounding test record and functional results as well.
Check how they protect functionality while improving speed
Ask for a recoverable backup, a rollback route, and an agreed deployment window. Staging is helpful where appropriate, but it does not reproduce every live integration or traffic condition. Someone still needs to check the live result after the approved deployment.
Write down the actions that must work: mobile navigation, form validation and delivery, consent choices, and relevant tracking. For stores, add product selection, cart updates, shipping calculations, and an agreed checkout test. Decide who supplies test accounts and who verifies the business-side outcome.
Be specific about “tested.” Seeing a form is different from submitting it and confirming receipt. Seeing a payment button is different from completing an authorized test flow. Testing should respect consent settings and avoid unwanted charges or customer notifications.
Ask how required scripts and styles will be protected when delaying JavaScript or removing assets. A feature that nobody exercised during testing can still be necessary. Disabling it may produce a cleaner report and a rather less useful website.
The proposal should identify who can reverse a problematic change and how you report it. That is especially relevant when the specialist, host, and existing developer each control different parts of the site.
Get scope, ownership, and support in writing
Clarify whether you are buying an audit, implementation, or both. Ask who makes the changes, which templates are included, and whether custom development, hosting work, or third-party troubleshooting requires a separate decision.
The handover should identify changed settings and code, retained exclusions, test results, unresolved issues, and recommended next steps. Another developer should be able to understand the work without reconstructing it from an invoice that says “speed.”
Clarify who controls plugin licenses and renewal accounts, what usage rights apply to custom code, and what remains available after the engagement. An included plugin is not automatically a lifetime subscription. Record ongoing costs before they become next year’s surprise.
Support for delivered changes also differs from ongoing maintenance. Agree on its duration, covered issues, response expectations, and contact route. Future redesigns, new tracking tags, and unrelated plugin updates may need separate work.
For the financial comparison, see my guide to WordPress speed optimization cost and the current WordPress speed optimization packages. A small, clearly defined job can be good value. A higher price needs a better explanation than a more expensive adjective.
Questions to send before hiring
Copy these into your enquiry and add your URL, important pages, symptoms, and previous attempts. The answers should help you compare the actual job.
1. Based on these symptoms, what would you investigate first, and what needs access before you can assess it?
2. Can you show a relevant example and explain your own contribution, testing, and remaining limitations?
3. Which pages and customer journeys will the quoted work cover?
4. What baseline, targets, testing conditions, and before-and-after evidence will we agree?
5. How will you protect the site, deploy changes, and test the functions we depend on?
6. What cannot you reasonably promise, and what might require extra approval or cost?
7. What will I receive at handover, including settings, code, license details, and unresolved issues?
8. If an optimization causes a problem, who handles it, within what support period, and under which terms?
My verdict
A narrow fix is enough when the problem is understood, the affected area is limited, and the result can be checked. You do not need a grand investigation to correct one clearly identified image-loading mistake.
Broader expertise makes sense when delays cross several templates, logged-in journeys, integrations, or server and browser work. In that situation, pay attention to diagnosis and testing before comparing the promises.
I sell this service, so apply these criteria to me too. My service page describes the process and deliverables; the scope still needs to fit your website. If you would like to discuss your website’s performance, send the URL and explain what feels slow.
Choose a WordPress speed optimization expert whose work you can examine after the screenshot has been saved. The score can look excellent. The checkout should still work.
Frequently asked questions
Can an expert guarantee 100/100?
A provider may offer a guarantee under defined test conditions. Ask which URLs, device profiles, tools, exclusions, and remedies it covers. It cannot establish perfect performance for every visitor or guarantee business outcomes.
Is a plugin setup enough?
It can be enough for a clearly identified configuration problem. Ask what the setup addresses and how its effect will be verified. Problems outside that scope need further investigation, not simply more enabled settings.
What proof should I ask for before hiring?
Request one relevant example showing the provider’s contribution and test context, plus a proposed measurement plan for your site. Treat a cropped score screenshot as limited evidence, not proof of dishonesty.
Can optimization break checkout?
Yes. Changes affecting required scripts, styling, or cache behaviour can interfere with store functions. Agree functional tests and a rollback process before implementation; a successful page-load test is insufficient.
What should the final report include?
Ask for comparable results, test conditions, functional-test outcomes, changed settings and code, exclusions, unresolved limitations, and support details. It should explain both what improved and what still needs attention.
Founder of WPservice.pro
Dalibor is a master of web excellence. With a Bachelor of Science (BS) in civil engineering, Dalibor had an unusual road to end up in IT. Cultivating deep expertise in WordPress website speed optimization, meticulous maintenance, development, and search engine optimization (SEO) while preserving his engineering approach to problem-solving.
Having completed over 90 projects and achieved a top-rated status (on Upwork) in the highly competitive digital niche, Dalibor is a proper authority on enhancing performance and ensuring websites look exceptional and perform flawlessly.
Dalibor is a published writer and an avid learner who continually explores and embraces the latest digital trends. With a commitment to quality and a keen eye for detail, Dalibor is your trusted guide to achieving web success.