← Blog

Fix a slow website before a sales campaign: who to hire

For e-commerce and operations managers with a campaign date and a leaky store: who to hire, in what order, and a 14, 7 and 3 day plan to protect checkout.

Get a summary withChatGPT Claude Google AI Mode
Calendar with a lit launch date and a glowing checkout path, a plan to fix a slow website before a sales campaign
On this page

Your campaign date is set, the store feels slow and shoppers leave at checkout. To fix a slow website before a sales campaign, hire a platform specialist for your store (Shopify, WooCommerce, Webflow or a custom build) plus a performance developer, and bring in conversion work only after speed. Allow about two weeks for a safe fix.

Less time still helps. Three days will not rebuild a theme, but it can remove the heaviest scripts, repair the worst mobile step and give you a tested way back if something breaks. The order of work matters more than the number of hours.

This guide is for e-commerce and operations managers who own the date and not the code. You get a way to find the leak, a plan for 14, 7 and 3 days out, fast wins for each platform, a symptom table and a checklist to vet whoever you hire.

Key takeaways

  • Hire a platform specialist and a performance developer first, and leave conversion work until speed and errors are fixed.
  • Allow about two weeks for a safe fix, and cut the scope to match if you have 7 days or 3.
  • Find the leak in your GA4 checkout funnel by device before anyone touches code.
  • Agree a freeze date, a staging copy and a rollback plan before the first change goes live.

Who to hire to fix a slow website

You need two roles, not one hero: a specialist who knows your platform and a developer who measures speed. A rush fix is a clean hand-off between them, with your campaign date at the end.

The two specialists to hire first

The platform specialist knows what is safe to change on your stack: Shopify Liquid themes and apps, WooCommerce plugins, Webflow's CMS and custom code, or your own Next.js application. They know which change breaks a payment step. The performance developer reads real visitor data, finds what holds the page back and tests every change before it goes live.

On a small store, one agency can supply both roles. On a larger one, expect two people and ask who coordinates them. A freelancer suits a contained fix with a single owner. An agency suits a store that needs both skills, a staging process and someone on call at launch.

Two hands passing a glowing baton toward a lit launch date tile, showing the hand-off between two specialists
Two specialists, one hand-off, one date

Why conversion work comes after speed

Conversion work changes what shoppers see: button copy, layout, offers. It needs steady pages and enough traffic to read a test, and a slow page adds noise to every result. Speed fixes remove weight and errors, so they rarely need a debate.

Baymard Institute documents an average cart abandonment rate (opens in a new tab) of roughly 70%, with causes such as unexpected costs and a long checkout. Speed and errors are the causes a developer can fix in days, which is why they come first. After the campaign you will have the traffic to test the rest.

Success measure

Ask for checkout completion on mobile to be the measure of success, not a speed score. A page can score well in a lab test and still lose carts at the payment step.

The minimum lead time, and what fits in less

Two weeks is the comfortable minimum. It fits a freeze, a measured baseline, two rounds of fixes on a staging copy and a load test. Shorter windows still work if you cut the scope to match:

  • 14 days or more: a full triage, fixes in two batches and a load test with time to spare.
  • 7 days: one round of fixes aimed at the single biggest leak, plus a rollback plan.
  • 3 days: removal only, such as dropping heavy scripts and repairing one broken step, with testing and a way back.
  • Under 3 days: a freeze, a monitored launch and a short list of fixes for after the campaign.

If the window is shorter than the work, start the campaign with your email list first. It gives you a live test before paid traffic arrives.

Find where your checkout loses sales

A developer fixes faster when you point at the right page. One day with data you already own shows whether shoppers leave at the product page, the cart or payment, and on which device.

Two documented numbers help you read it. Google's Core Web Vitals (opens in a new tab) set the line for a good experience, measured at the 75th percentile of real visits. Baymard gives the typical abandonment to compare against. Your own number matters more than either.

2.5s

LCP target

How fast the main content appears. Google's good threshold, 75th percentile

200ms

INP target

How fast the page answers a tap. Google's good threshold

~70%

Average cart abandonment

Baymard Institute benchmark, not your store's number

Google also sets 0.1 as the good line for CLS, how much the layout jumps while the page loads. Jumping layouts matter at checkout, because a button that moves under a thumb costs an order.

Translucent funnel with a small leaking crack under a magnifier, showing how to find the leak in a checkout
One leak in the funnel is enough to find first

Read the GA4 checkout funnel by device

In GA4, open Explore and build a funnel from the standard e-commerce events: view_item``add_to_cart``begin_checkout``add_shipping_info``add_payment_info and purchase. Break it down by device category and compare the last four weeks. Using the same weekdays keeps weekend noise out.

Look for a step where mobile loses far more shoppers than desktop. That step is your first lead. If an event is missing, the funnel shows a false cliff, so check that each one fires before you trust the picture.

Here is an illustrative example. A store expects 40,000 checkout starts in campaign week and completes 10%, so 4,000 orders. One more point of completion is 400 orders, and at an average order of $60 that is $24,000. One repaired step beats ten small tweaks.

Watch Clarity recordings of abandoned carts

Microsoft Clarity is a free session recording tool. Filter the recordings to mobile, to the cart and checkout pages, and to sessions that ended without a purchase. Watch about twenty.

Look for taps that do nothing, a spinner that never ends, fields that jump and shoppers who go back to edit their details. Write down the step and the time for each one. These notes become your developer's brief.

Check Core Web Vitals field data

Field data comes from real visitors, not from a test on your laptop. Search Console's Core Web Vitals report and PageSpeed Insights show it by page group, so check the product, cart and checkout templates on mobile. Google's page experience documentation (opens in a new tab) explains how these signals relate to search.

On Shopify, the checkout is largely hosted by Shopify, so the leak is often in the pages and scripts before it. If the data shows many causes at once, a performance and security audit lists them and ranks them by effort.

Your countdown to fix a slow website

Each window answers one question: what can we change safely by the date? The order never changes: freeze, measure, fix, test, launch.

Campaign countdown

  1. 01Freeze14 days: no new apps, line drawn
  2. 02Measure14 days: funnel, recordings, field data
  3. 03Fix7 days: top two leaks on staging
  4. 04Load test3 days: simulated traffic, rollback ready
  5. 05Monitored launchLaunch: checkout completion holds

Five-step countdown to launch: freeze, measure, fix, load test and monitored launch, with 14, 7, 3 and 0 days

The countdown from freeze to a monitored launch

The four phases below show what each window holds. The sections after them give the detail you can hand to your team.

  1. 14 days out

    Freeze and measure

    No new apps, plugins or theme changes. Record the baseline funnel by device.

  2. 7 days out

    Fix the biggest leaks

    Release the top two fixes from staging, one at a time.

  3. 3 days out

    Load test and rollback

    Simulate campaign traffic, place test orders and name who can roll back.

  4. Launch day

    Monitor the first hours

    Watch live checkout completion with the way back ready.

14 days out: freeze and measure

Freeze first. From today, no new apps, plugins, theme edits or banners except the campaign's own, because each change makes the data harder to read. Then make a staging copy, a private duplicate of your store where fixes are tested before shoppers see them.

Record a baseline: the funnel by device, field data for the three templates, page weight, error logs and the time each checkout step takes. Write down the success measure, checkout completion on mobile. Finally, rank the leaks by revenue at stake against effort.

7 days out: fix the biggest leaks

Work on the top two leaks only. Typical fixes at this stage are removing or delaying heavy scripts, shrinking images, repairing a payment error and caching what can be cached. Release in small batches and check the funnel after each one, so you know which change helped.

Close the fix list around day 4. Anything that does not fit goes onto an after-campaign list with a name and a date, so nobody improvises in the last days.

3 days out: load test and prepare rollback

A load test simulates many shoppers at once. Take your last campaign's peak of concurrent visitors, add a margin and run it on product, cart and checkout pages. On a hosted platform such as Shopify, the platform handles the traffic, so you test your theme, apps and third-party calls. Check its rules before sending synthetic traffic.

Then place test orders on every payment method, on mobile, with a discount code and each shipping option. Prepare the way back: a duplicate of the live theme or a full backup, a list of plugin versions and one named person with permission to roll back. Agree the trigger in advance, for example checkout completion falling clearly below baseline for thirty minutes.

Launch day: monitor the first hours

Launch with someone watching. Check live checkout events in GA4, the error log, the payment provider's dashboard and the support inbox at launch, after one hour and after four hours. If the agreed trigger fires, roll back first and investigate after. Keep the freeze in place until the campaign ends.

Fast wins by platform

These changes tend to pay back within days and carry low risk. Each one starts in plain words, so you can find your stack and ask your specialist to begin there.

Shopify: app and script inventory

List every installed app and every script in the theme and the tag manager. Many apps add JavaScript to every page, even where they do nothing. Remove the apps nobody uses, then delete the leftover code they leave in the theme.

Next, delay what can wait. A chat widget or a review carousel can load after the shopper's first tap. Shopify's theme performance best practices (opens in a new tab) cover app and script weight and image handling. If you run a large catalog with many integrations, see how to choose a Shopify Plus agency.

WooCommerce: plugin and query cleanup

Deactivate plugins that nobody uses and check which ones load scripts on the cart and checkout. Use a query monitor to find slow database queries, often caused by a plugin or a large product filter. Confirm that caching excludes the cart and checkout, since a cached cart shows the wrong totals.

Deep checkout tuning is a longer job than a countdown allows. If your store needs more than a cleanup, read how to choose a WooCommerce development agency for a mid-sized store.

Webflow: image and embed weight

Webflow's hosting and CDN are solid, so weight usually comes from what you add. Serve images at the size they display and in a modern format such as WebP. Replace background videos with a still image on mobile. Load embeds like reviews, maps and chat only on interaction, and remove custom code that no longer does anything.

Custom builds: caching and third-party calls

On a custom application, the slow step is often a call to another system: ERP stock, shipping rates or a tax service. Cache what changes rarely, set timeouts with a fallback, and never make the payment step wait for an analytics call. Check database indexes on order and product queries, and serve images from a CDN.

Match each symptom to its fix

Use this table to brief your developer and to judge a quote. The effort ranges are typical for a focused fix by an experienced developer. Your stack and code can move them.

SymptomLikely causeFixTypical effort
Product pages load slowly on mobileOversized images, many scriptsResize images, delay non-essential scriptsHalf a day to 1 day
Cart page spins or freezesHeavy app or plugin, slow shipping-rate lookupRemove or delay the add-on, cache rate calls1 to 2 days
Layout jumps while shoppers tapImages without set sizes, late bannersReserve space for images and bannersHalf a day to 1 day
Payment step fails or times outGateway script conflict, server errorsUpdate the gateway, remove conflicting scripts, read the logs1 to 3 days
Buttons answer lateHeavy JavaScript from chat, tags and popupsRemove or delay what is not needed1 to 2 days
Site slows as traffic risesHosting limits, uncached pages, slow queriesAdd caching, fix queries, resize hosting2 to 5 days
Slow for shoppers far from the serverNo CDN, distant originAdd a CDN for static files1 to 2 days
Checkout fine in tests, poor for usersThird-party tags in real browsersCompare field data, remove the tagsAbout 1 day

If a quote lists work that matches none of these rows, ask which data it came from.

A hiring checklist for a rush fix

You have days, not weeks, to choose, so use questions that reveal how someone works. Our longer list of questions to ask a web development agency covers the wider hire.

Questions to ask before you sign

  • Who does the work, and are they available every day until launch?
  • Which metric proves success, and will you report it by device?
  • What have you fixed on my platform before, and what would you change first?
  • How do you test, and what is the rollback if checkout completion falls?
  • What is out of scope, and where does the work stop?

Red flags in a rush quote

  • A promise of a speed score instead of checkout completion. A score can rise while carts still fail.
  • Changes made straight on the live store.
  • A quote sent before anyone looked at your data.
  • A redesign offered as the fix a few days before launch.
  • No named person and no reply time.

Staging, rollback and who is on call

Staging and rollback are the two things worth insisting on. Staging lets a change fail where shoppers cannot see it. Rollback turns a bad release into a ten-minute problem. Ask who is reachable on launch day and through which channel.

  • Name one person who does the work and is reachable daily
  • Get a staging copy of the store before any change
  • Agree a written rollback plan and who can trigger it
  • Agree the metric: checkout completion by device
  • Set the freeze date and share it with the whole team
  • Give the developer access to GA4, Search Console and the theme or code
  • Put the scope in writing with a clear stop line

What werun.dev does when the date is close

You set the goal: a campaign that converts. werun.dev builds the way there as a short, focused project that follows the order in this article. We have looked after 200+ sites over 12+ years, and a pre-campaign speed fix has a page of its own.

Audit in the first days

We read your funnel by device, your recordings and your field data. Then we review apps, plugins and scripts on a staging copy. You receive a short written list ranked by revenue at stake and effort.

A scoped fix with a clear stop line

We agree the fixes in writing, release them in small batches and stop when the scope is done. What does not fit goes onto an after-campaign list. You follow progress in the client panel and can ask for a change by clicking on the page with Client Feedback.

A monitored launch

On launch day we watch checkout completion with the rollback ready. We reply in under 2 hours on business days. The work can continue after the campaign on a monthly Retainer or Care plan if you want it.

Where to start this week

Begin with what costs nothing. Four steps are enough to turn a worry into a plan:

  1. Pull the GA4 checkout funnel by device for the last four weeks.
  2. Set the freeze date, 14 days before launch if you can.
  3. List every app, plugin and script on the store.
  4. Write the success measure: checkout completion on mobile.

Then tell werun.dev your campaign date and goal. We reply in under 2 hours on business days and tell you what fits before the date.

Frequently asked questions

How long does it take to fix a slow website before a sales campaign?

About two weeks gives time to freeze changes, measure, fix on a staging copy and load test. With 7 days you can fix the biggest leak, and with 3 days you can remove heavy scripts and prepare a rollback.

Who can fix a slow ecommerce site fast?

A platform specialist for your store (Shopify, WooCommerce, Webflow or custom) working with a performance developer. An agency that offers both roles usually moves faster than coordinating two separate freelancers.

Should I hire a freelancer or an agency to fix my store before a campaign?

A freelancer fits a contained fix with one owner and a clear scope. An agency fits when you need platform and performance skills, a staging process and someone on call on launch day.

How do I find out why my checkout is losing sales?

Build a checkout funnel by device in GA4, watch Clarity recordings of abandoned mobile carts and check Core Web Vitals field data for product, cart and checkout pages. Together they show the step, the device and the likely cause.

Can I reduce checkout abandonment in a week?

Part of it, often. Removing heavy scripts, repairing a failing payment step and fixing layout jumps on mobile can be done in days, but a checkout redesign cannot. Measure completion by device before and after so you know what worked.

Should I delay my campaign if my store is still slow?

Delay only if checkout fails for many shoppers and the fix cannot be tested in time. Otherwise, launch to a smaller audience such as your email list first, keep a rollback ready and watch completion in the first hours.

Running into this on your own stack?

Tell us the platform and the problem. We'll tell you what it takes and who does it.

Start a conversation No commitment. Just a conversation.