If your website never loaded Optimizely's testing script, this attack did not reach your visitors through that script. A fix going forward does not tell you what your pages were doing before it, so it is still worth checking. The wider lesson is the uncomfortable one: a website can be compromised through a third-party dependency without its own code or infrastructure ever being breached. Below is what happened in plain words, a quick way to list every outside script your site loads, how to spot the ones that can see payment or patient forms, and when it makes sense to have someone lock it down.
What happened with Optimizely, in plain words
Optimizely sells tools that let marketing teams test different versions of a page and see which one works better. To do that, a piece of the vendor's code runs on the customer's website, inside each visitor's browser. Optimizely's own press release states that over 10,000 businesses deliver digital experiences with Optimizely.
On 16 November 2026, malicious JavaScript was served to visitors of 178 organisations through the Optimizely Web experimentation platform, according to the researchers' analysis. Optimizely said on its status page that an unauthorized party exploited a vulnerability in Experimentation and used it to add malicious JavaScript to some customer projects.
It is worth looking at where the code was placed. The Optimizely snippet contains Optimizely's code, the customer's experiments and an optional Project JavaScript field: custom code the customer writes in their Optimizely dashboard, which runs on every page.
The scale, as measured by the researchers who found it:
Researchers checked 1,611 Optimizely snippets and found 188 carrying the malware. Banks and financial firms were the largest group of victims, with 32 organizations. These are companies with large security teams, which is exactly why a small business should not assume "it couldn't happen to us."
What the malicious script actually did
This was not a prank or a defaced homepage. It was built to steal:
- Card numbers. The script watched form submissions and every fetch/XHR POST for a card number (Luhn-checked), expiry date and CVV. Put simply, it looked for a card number that passes the Luhn check and has a known brand prefix, a valid expiry date and a three- or four-digit CVV. In other words, it was looking specifically for card details that look genuine.
- Logins and sessions. The malware skimmed payment cards and stole cookies and screenshots. Cookies are what keep a visitor logged in, so a stolen cookie can be as useful to an attacker as a stolen password.
- Crypto payments. It swapped cryptocurrency wallet addresses for the attacker's.
- iPhones and iPads. On iPhones and iPads the script read the iOS version and loaded an exploit chain in a hidden iframe, and it delivered the Coruna and DarkSword iOS exploit kits. That means a visit to a trusted site could put the visitor's phone itself at risk, not just the form they were filling in.
The site owners did nothing wrong in the usual sense. They did not get phished and they did not skip a plugin update. They trusted a vendor, and the vendor's delivery pipe was poisoned. That is why the question "is my website affected?" has two answers: the narrow one about this vendor, and the broader one about every outside script you load.
Has this vendor been hit before?
SecurityWeek has reported a different kind of incident at the same company: a threat actor gained access to Optimizely's systems through a sophisticated voice-phishing attack. That incident was confined to certain internal business systems including Zendesk, records in Salesforce CRM, and a limited set of internal documents used for back-office operations. Ask the vendor whether any earlier incident involved customer data or personal information, and get its answer in writing. The two events are not the same, but together they make a point every owner should hear: a vendor's security is now part of your security, whether you signed up for that or not.
Is your website affected? The two-minute answer
Ask yourself, or whoever built your site, these questions:
- Do we use Optimizely Web (or "Optimizely Experimentation") on the site? Your developer, agency or marketing lead should be able to answer this quickly. If nobody can, that is useful information in itself.
- If yes, was it on any page with a checkout, login, booking or intake form?
- Did customers enter card details or log in on the site while the malicious code was live? If so, talk to your payment processor and your IT or legal contact today. Do not wait for a customer complaint.
If the answer to the first question is "no" or "I don't know", keep reading. In our experience "I don't know" is a common answer, and it is the real problem. You cannot judge your risk from a news story if you do not know what your own pages load.
You do not need to be technical to start. You need a laptop with a desktop browser, your website address, and a little time. Write things down as you go.
How to list every outside script your site loads in 15 minutes
This is our own simple method for owners and office managers. It will not catch everything a professional audit would, but it gives you a list you can act on right away.
Minutes 0-3: Pick the pages that matter
Do not try to check the whole site. Pick three to five pages:
- Your homepage
- Any page where people pay (checkout, "pay your invoice", deposit page)
- Any page where people type sensitive details (new-patient intake, appointment booking, case intake, quote request, login)
- The thank-you or confirmation page after a form
Minutes 3-10: Look at what each page loads
On a desktop browser:
- Open the page, right-click anywhere, and choose Inspect.
- Click the Network tab, then reload the page.
- Click the filter labelled JS (or "Script"). You now see the scripts the page pulled in.
- Look at the Domain column (right-click the column headers to switch it on if needed). Anything that is not your own domain is an outside script.
- Copy each outside domain into a simple list or spreadsheet.
Repeat for each page. Then, if your site uses Google Tag Manager or a similar tag manager, log in and list every tag in there too. In our experience tag managers are a common blind spot, because they let anyone with access add scripts without touching the website itself.
Minutes 10-15: Name each one and write down who owns it
For every domain on your list, write three things next to it: what the tool is (chat, booking, analytics, ads, A/B testing, reviews, payments), who in your business asked for it, and whether you still use it. When we do this with clients, we usually find at least one script nobody remembers adding, often from an old marketing campaign or a former agency.
Do not delete scripts from a live site on a guess. Some are load-bearing (your payment form or booking widget may be one). Mark them for review first, and remove them only in a planned change that you test afterwards.
Which outside scripts can see payment or patient forms?
Here is the part that is easy to overlook. A script that runs on a page can generally read what is typed into that page. So it helps to ask not only whether a vendor is trustworthy, but which pages its script runs on and what people type there.
Here is how we sort the usual suspects on a small business site:
| Script type | Common examples | Typical placement | Risk to forms |
|---|---|---|---|
| A/B testing / personalisation | Optimizely, similar tools | Often every page | High โ runs widely and is designed to change pages |
| Tag managers | Google Tag Manager | Often every page | High โ can load anything, added by anyone with access |
| Chat widgets | Live chat, AI chat bots | Often every page | Medium-high โ on the same page as forms |
| Analytics and session replay | Analytics, heatmaps, screen recorders | Often every page | Medium-high โ worth checking whether it records form input |
| Ad and social pixels | Ad network pixels | Often every page | Medium โ sit beside forms |
| Booking / intake widgets | Online booking, patient intake | Booking and intake pages | High โ they are the form |
| Payment fields | Stripe, Square, processor checkout | Checkout only | Lower if hosted in the processor's own frame or page |
| Fonts, maps, reviews | Font services, map embeds, review badges | Various | Low-medium |
The pattern we look for: tools that run on every page, including your checkout or intake page, deserve the closest attention.
A worked example
Hypothetical: a small physical therapy clinic in the San Fernando Valley runs its site on a page builder. The new-patient form asks for name, date of birth, insurance details and a short medical history. A 15-minute check finds nine outside scripts on that page: analytics, a session-recording heatmap added two years ago, two ad pixels, a chat bubble, a review badge, a font service, the booking widget, and a tag manager that loads three more tags.
Of those, the booking widget is needed. The heatmap is no longer used and may record what patients type. The ad pixels and chat bubble do not need to be on the intake page at all. The fix is not dramatic: remove the heatmap, keep marketing tags off the intake and payment pages, lock down the tag manager to one named person, and add a browser rule (more on that below) that only lets the intake page send data to the booking vendor. The clinic goes from nine outside scripts with access to patient details to one or two.
What it could cost a California business
We cannot tell you what a breach would cost your business, and nobody honest can. But some published figures are worth knowing if you do business here. Check with your lawyer whether and how these apply to you.
Privacy law. These amounts went into effect on January 1, 2025. A skimmer that runs on a busy booking page can touch a lot of consumers, so per-consumer and per-violation figures matter.
Card rules. If you take card payments on your own site, the payment card industry standard matters. The PCI Security Standards Council wrote that of the 64 new requirements in PCI DSS v4.0, 51 are future-dated and will be effective as of 31 March 2025. Ask your payment processor, or the person who handles your compliance questionnaire, which requirements apply to you, and ask specifically how the scripts on your payment page are tracked and controlled.
The wider fraud picture. More than one million complaints were submitted during the year, with fraud accounting for the majority of losses. Take stock of which off-the-shelf widgets, such as scheduling, payment, chat or ordering tools, sit on your own pages.
Beyond any fines there is a cost you cannot easily measure: the trust of the people who paid you or handed you their details.
How to lock it down
You do not need to rip out every useful tool. Aim for fewer scripts on sensitive pages and a browser-level rule about what is allowed to run and where data may go.
1. Cut the list
Remove anything no one owns or uses. Keep marketing, testing and chat tools off checkout, login and intake pages unless there is a strong business reason. In our view this is the cheapest step, and the one to start with.
2. Move card entry to your processor
Where possible, let your payment processor host the card fields, either as an embedded frame from their domain or a hosted checkout page. That way, scripts on your own page are kept further away from the card number. Ask your web developer whether your checkout works this way today.
3. Add a Content Security Policy with reporting
Put simply, this is a short set of rules your site sends to the browser, saying which scripts may run and where the page may send data. In its write-up of this attack, Report URI advises deploying a CSP, restricting connect-src, avoiding unsafe-eval where you can, and turning on reporting. Even a partial policy can help, because restricting where a page can send data limits what a malicious script can do with card, cookie or screenshot data. Reporting matters as much as blocking, because it means the browser tells you when something unexpected tries to run or send data, instead of you finding out from a customer.
A sensible order is to start in "report-only" mode, watch the reports for a while, then switch to enforcing.
4. Pin scripts that should never change
Subresource Integrity (SRI) lets browsers verify that fetched resources are delivered without unexpected manipulation, by supplying a cryptographic hash the resource must match. If the file is modified, the hash won't match and supporting browsers will refuse to load it. Use it for fixed library files you host or pin to a specific version. For tools whose vendor updates the code on their side, ask your developer whether integrity pinning is workable; if it is not, lean harder on limiting which pages those tools run on and where those pages can send data.
5. Treat vendors as part of your supply chain
CISA's recent guide for software customers covers procurement, testing, deployment and patching, and includes guidance on software bills of materials. A small business does not need the full process, but the idea scales down: keep your script list, review it every quarter, and ask each vendor how they would tell you if they were compromised.
Common mistakes we see
- Assuming a big platform covers you. Your hosting platform looks after its own servers. You still decide which widgets get pasted into your pages.
- Leaving old agency tags in place. We regularly find former agencies and freelancers who still have tag manager access long after the contract ended.
- Turning on a strict CSP in one go. This can break your booking or payment flow on a busy day. Start in report-only mode instead.
- Assuming "fixed by the vendor" means "nothing to do". Optimizely says it blocked access to the vulnerable path, deployed a fix, and removed the injected code from affected projects. That helps going forward, but it is still worth asking what your pages did before the fix. If you used the tool on payment or login pages, talk to your processor and consider prompting customers to watch their statements.
- Looking at the homepage alone. The pages that matter are checkout, login and intake.
What we'd do for a Los Angeles shop or clinic
When a business asks us to look at this, our website security and web development work usually follows the same short path:
- Inventory. A full list of every outside script, tag and pixel on every page that handles money or personal details, with an owner next to each.
- Trim. Remove the dead ones, move marketing tools off sensitive pages, and tidy tag manager access.
- Protect. Add a browser script policy in report-only mode, tune it against real traffic, then enforce it. Pin static libraries with integrity checks. Confirm card entry is hosted by the processor.
- Watch. Set up reporting so an unexpected script triggers an alert, not a customer complaint. If you want to see how we approach ongoing monitoring, our Mal-Mon malware monitoring case study shows the kind of tooling we build.
We work with businesses across the region, from Santa Monica to the San Gabriel Valley; see our Los Angeles web services page for more on how we work locally. In our experience, sites on the common website builders and content systems can usually be tightened without a rebuild.
When to call someone instead of doing it yourself
Do the quick inventory yourself. Bring in help when any of these is true:
- Your site takes card payments directly, or collects health, legal or financial details.
- You found scripts nobody can explain.
- You used Optimizely Web or a similar testing tool on a checkout, login or intake page.
- You want a Content Security Policy but cannot risk breaking bookings or payments.
Next step
Start with the checklist above today. It costs nothing, and it shows you what your own pages are really loading. If the list is longer than you expected, or it includes anything on a payment or patient page, book a Free Automation Audit and we will review your script inventory with you, tell you plainly what is risky, and show you what locking it down would involve.
Sources
- CISA โ Securing the Software Supply Chain: Recommended Practices Guide for Customers (November 5, 2026)
- California Privacy Protection Agency โ California Privacy Protection Agency Announces 2025 Increases for CCPA Fines and Penalties (December 17, 2024)
- Optimizely โ Optimizely Reaches $400M ARR Milestone as Demand for Marketing Operation System Surges (May 14, 2024)
- Optimizely Status โ Optimizely Web Experimentation Security Incident (November 17, 2026)
- PCI Security Standards Council โ Now is the Time for Organizations to Adopt the Future-Dated Requirements of PCI DSS v4.x (August 20, 2024)
- Report URI โ World's Largest Brands Infected With Malware Via Optimizely (November 16, 2026)
- StepSecurity โ OPTIMJACK: Optimizely-Hosted Scripts Loaded a Crypto Clipper and Card Skimmer on 178 Organizations (November 17, 2026)
- Help Net Security โ Malware spread through Optimizely hits 178 organizations, including BBC, CNN and Zoom (November 17, 2026)
- Help Net Security โ Cybercrime losses break the $20 billion mark (April 7, 2026)
- SecurityWeek โ Optimizely Vulnerability Exploited to Serve Malware on Major Websites (November 17, 2026)
- SecurityWeek โ Ad Tech Company Optimizely Confirms Cyberattack (February 25, 2026)
- MDN Web Docs โ Subresource Integrity (SRI) implementation guide (December 29, 2025)