Black Friday is eight weeks away, and clients are starting to ask whether their stores are ready. The answer depends on what “ready” means, and that was the starting point for today’s webinar.
Peak season, every season: Building a repeatable Black Friday Cyber Monday playbook was hosted by Jeff Golenski, Product Manager at Automattic for Agencies. He was joined by Steven Sauders, Senior Solutions Engineer at Pressable, and Brent MacKinnon, who leads Product Marketing and Developer Advocacy at WooCommerce. Brian Coords, a Developer Advocate at WooCommerce, ran a live demo.
Start eight to 10 weeks out, with the business
Brent’s advice is to talk about the business before you touch the site. Agree with your client on:
- What’s on sale, and how big the discount is.
- How many visitors and orders to expect.
- How much stock is available.
- How many orders the team can realistically pack and ship.
Three to four weeks out, check the site: hosting, payments, and every plugin or service the store depends on. For the final two weeks, stop making changes and retest everything. As Brent said, if there’s not enough time to test a specific change before it goes live, it’s probably too late for that change.
Your countdown to Black Friday
10
Eight to 10 weeks out
Agree on the business plan
4
Three to four weeks out
Check hosting, payments, and plugins
2
Final two weeks
Stop making changes and retest
BF
Sale day
Watch your metrics closely
Test the full customer journey
The step agencies most often leave too late is the end-to-end test. Follow the path a real customer takes. They see an email or ad, click through to the sale page, add a product to the cart, apply the discount, pay, and get a confirmation email. Then check that the order reaches the people who pack and ship it.
The customer journey
Email or ad → Sale page → Cart → Discount → Payment → Confirmation → Fulfillment
That full run shows you the gaps between marketing, the store, the payment provider, and fulfillment. It also tells you whether you know who can pause an ad, fix a broken discount, or tell customers about a delay.
Sort your client list by risk
If you manage dozens of client stores, you can’t audit every one in depth. Before joining Pressable, Steven ran an ecommerce agency with about 150 clients. He sorts stores using three signals:
Signal 1
Your gut
You already know which sites give you trouble.
Signal 2
Hosting metrics
A site that already uses most of its CPU or PHP workers on a normal day needs attention before peak traffic arrives.
Signal 3
Expected growth
A store expecting one to two times its usual checkouts carries less risk than one expecting five to 10 times.
1–2× lower risk 5–10× higher risk
His reusable audit covers CPU and PHP worker usage, caching, and an incident plan. Caching matters because WooCommerce can’t cache the cart or checkout. Caching every other page leaves more server capacity for the cart and checkout.
Plan for problems before the sale
Steven and Brent both recommend deciding a few things ahead of time, and working through the order step by step when something breaks on the day.
Checklist
Decide before the sale
☐ What would make you restore a backup
☐ Whether you’ll push a fix for a small bug during the sale
☐ Who investigates problems, issues refunds, contacts customers, and talks to support
Troubleshooting
When an order breaks on the day
- Was the customer charged?
- Did WooCommerce record the order?
- Is it stuck as pending, on hold, or failed?
- Did the email go out, and did the stock count change?
- Compare WooCommerce with your payment provider before you ask anyone to pay again.
On the hosting side, Pressable gives you burst capacity for unexpected spikes, a failover server, and a support team on hand during the sale. Jetpack real-time backups let you roll back a bad change and keep every order placed since then. Jetpack is included free with Pressable through Automattic for Agencies.
How Claude found the bottleneck hiding in a loyalty program
Steven’s demo store, a dog food shop called Kibble Standard, handles 10 checkouts per minute and expects 40 on Black Friday. Using Pressable MCP, Claude pulled the site’s metrics and found it was already at its PHP worker limit on a normal day. At four times the traffic, shoppers would be stuck waiting at checkout.
Claude traced the slowdown to the store’s loyalty program. It ran the same database query four or five times on every cart load, against a table of 3.8 million rows. A single database index cut cart load time from about 6.5 seconds to 1.3 seconds. Claude also suggested skipping the loyalty lookup for logged-out shoppers, who can’t earn points anyway.
Claude applied and tested the fixes on a staging site so Steven could review them before they reached the live store. It also built a dashboard for monitoring the store during the sale.
A Black Friday readiness report from one prompt
Brian gave an AI agent a campaign brief, with a sweater discount, a coupon code, and sale dates, and asked whether the store was ready. The agent worked in the background while he did other work.
Its report flagged a missing coupon, abandoned cart recovery switched off in AutomateWoo, and a backlog of on-hold orders that could slow things down during a sales surge. It also noticed it had been pointed at a staging site. The agent had read-only access, so nothing on the store changed.
Brian then asked for a sale page. The agent built a draft from the store’s block theme patterns, with real product details and a call to action. It also reminded him that the coupon still needed to be created.
All of this runs through one MCP connection to the WordPress site. That connection gives the agent access to tools from WooCommerce, its extensions, and other plugins. Agencies can also write custom tools for each client as PHP functions.
Where to draw the line with AI
Both demos use MCP, an open standard, so the Pressable and WooCommerce tools work with any AI agent that supports MCP connections.
Both panelists use agents to run the same checks across many stores: spotting failed tasks, odd order activity, and payment problems, or drafting replies to customers. Neither lets an agent change a live store without approval. Brent wants someone to sign off on anything that touches pricing, discounts, products, plugins, or permissions. Steven reviews every code change before it goes live.
Your playbook at a glance
1
Eight to 10 weeks out
Agree on the offer, expected traffic, stock, and fulfillment.
2
Three to four weeks out
Audit hosting, payments, and plugins. Run a full customer journey test.
3
Two weeks out
Stop making changes, retest, and confirm who owns each part of the incident plan.
4
Sale day
Watch your metrics, and work through problems step by step before you act.
Watch the recording
The full recording is here. Not a partner yet? Automattic for Agencies is free to join.
