Peptide Clinic Marketing: What to Measure Weekly

Peptide clinic marketing cannot be measured the way an ad account is measured, because at most peptide clinics the ad account is the least dependable part of the mix. The question that matters every week is which of your messier channels (organic search, content, referrals, email, word of mouth, and whatever paid campaigns survived review) actually produced consult requests, and that question can only be answered from your own site data, never from a platform dashboard. Below is the weekly measurement set we would run at any clinic offering peptide therapy alongside hormone, TRT, longevity, or weight management services, and how Curve covers each item under a BAA, with Curve Analyst answering the questions in plain language from the clinic's own data.

TL;DR

• Paid channels are unreliable for peptide clinics because platforms treat many peptides as unapproved or restricted substances, and the status of individual peptides keeps changing. Plan for disapprovals and measure everything else properly.
• Count consult requests by traffic source every week, with organic, referral, email, and AI assistant sources on the same chart as paid.
• Treat each treatment page as its own campaign. Sessions and consult rate per page tell you what people came for and whether the page closed.
• Find the step between treatment page and consult booked that lost the most people, and fix that before touching a headline.
• Compare to last week before reacting, and confirm the conversions you are allowed to send are still leaving Curve.
• Keep the analytics aggregate and under a BAA. A visit to a page that names a sensitive treatment becomes health information the moment it is attached to an identifier.

Why a peptide clinic cannot be measured like an ad account

The ad platforms decide, week to week, whether your clinic is allowed to exist in their auction. Peptides sit in a contested regulatory space, the platforms classify many of them as restricted or unapproved substances, and the classification of any specific compound can change without notice. So campaigns get disapproved, accounts get restricted, and a clinic that budgeted for steady paid traffic gets a trickle instead. The mechanics of what you can and cannot say are covered in our guide to peptide and hormone clinic ad policy boundaries, and the parallel rules for HRT and TRT are in the hormone therapy advertising guide.

The consequence for measurement is bigger than most owners realize. With paid intermittent, the clinic leans on organic search, educational content, referrals from other practices and trainers, email to an existing list, and plain word of mouth. Each of those channels shows up differently in analytics, and none of them shows up in an ad platform at all. Google Ads knows about Google Ads. Meta knows about Meta. Neither of them has any idea that your longevity article was cited by an AI assistant on Tuesday and produced consult requests through Thursday.

So the weekly question changes. An aesthetic practice with a healthy paid program asks which ad won. A peptide clinic asks which channel produced consults at all, and whether that moved since last week.

The weekly measurement set for a peptide clinic

Consult requests by channel, including the channels ad dashboards ignore

Define a consult request as the goal, whether that is a form submission, a booking, or a tracked call, then ask how many fired last week by traffic source. Organic search, referral, email, direct, and paid belong on the same chart, because the whole point is comparison. A clinic that reads only its paid report is measuring the channel most likely to have been switched off mid-week.

Two sources deserve special handling. AI assistants increasingly send readers to peptide and longevity content, and Curve keeps the referring hostname as the source, so chatgpt.com, perplexity.ai, and similar appear as their own rows and Analyst can filter on them. At the channel level they currently count as referral; there is no dedicated AI assistant channel yet, so ask by source when you want them separated. Word of mouth is the other one. It lands in direct, with no clean way to split it out, so watch direct as a trend instead of trying to explain each visit.

Whatever conversion count you look at here is Curve's own attribution, credited by Curve's attribution engine using the model you chose. It will not match what a platform claims, and it should not, because the platform only sees its own clicks. Analyst says which figure it is quoting.

Treatment page traffic and consult rate

Your treatment pages are the real campaigns. A peptide clinic typically has pages grouped by category, such as recovery, metabolic, longevity, and cognitive support, often beside hormone and weight management pages, and each draws a different visitor with a different intent. Ask which pages had the most sessions and which had the highest consult rate, and read the two lists against each other.

They usually disagree. The page with the most traffic tends to attract researchers comparing compounds, and it converts poorly. A page with modest traffic and a high consult rate is telling you the people who reach it already decided, and it is often the page most worth building organic content around. Content pages belong in this view too: an article that sends readers on to a treatment page is doing work no ad platform will ever credit.

Our list of analytics questions clinics should start with covers the general version of this routine. The peptide version weights treatment pages more heavily, because that is where policy uncertainty and patient intent collide.

Where the funnel loses people between treatment page and consult booked

The peptide funnel is usually content or ad, then treatment page, then consult request, then consult completed, then protocol started. A configured funnel in Curve shows step-by-step drop-off for the on-site portion, from treatment page through consult request. The later steps happen in your scheduling and clinical systems and only appear in Curve if you configure them as goals, so decide early whether you want them reported beside the web steps.

Ask which step lost the most people. A large drop between the treatment page and the start of the consult form usually means the page answered the wrong question, buried the call to action, or made the reader nervous about what they were signing up for. A drop between form start and submission points at the form itself: too many fields, a request for information the visitor does not want to type before talking to anyone, or a mobile layout that breaks. A drop after submission is a scheduling and response problem; marketing cannot fix it but should flag it, because marketing paid for the visit.

What changed versus last week, and why

Every number above needs last week beside it before you react. Peptide clinics have their own reasons for week-over-week swings: a campaign disapproved on Wednesday, which shows up as paid falling while every other channel holds; a referral partner sending a batch of clients after an event; an email going out to the list; an article picked up by an AI assistant. Each has a signature in the channel breakdown, and the comparison is what lets you tell them apart.

Analyst does period comparisons on goals and treats unknown and zero as different things. That distinction matters here. A channel that reads zero because nothing arrived is a marketing finding. A channel that reads unknown because data is still syncing is a wait-and-see, and Analyst says which one it is before it gives you the number.

Whether tracking still delivers the conversions you are allowed to send

For a peptide clinic, the set of conversions you can send to an ad platform is narrower than for most businesses, because of policy and because of what a compliant setup strips before anything leaves. Curve receives events on its own US-hosted infrastructure first, removes what should not leave, and forwards the permitted conversions to Google Ads, Meta, and the others. Whatever you decided is permitted, you want to know each week that it is still arriving.

The reconciliation view in campaign reporting puts what Curve sent to Google or Meta next to what the platform credited. A gap is usually the platform's attribution window still being open rather than under-reporting, so do not panic at a small one. The failure that should worry you is a goal reading zero for the week while the form on the site still works, because that is tracking, not demand. Analyst does not yet report on tracking configuration or connector health, so this check is partly a Curve dashboard item; what Analyst will do is show the goal count dropping to nothing and say so plainly instead of smoothing it over.

Why aggregate, BAA-covered analytics matters more when your pages name the treatment

A retailer can be sloppy about analytics and mostly harm itself. A peptide clinic cannot, because a treatment page URL says what the visitor is considering for their body, and once it sits next to an identifier such as an IP address, a client ID, or a form field, you are holding information about an individual's health. Under HIPAA, any vendor handling that on your behalf is a business associate and needs a BAA. Beyond HIPAA, the FTC has pursued disclosures of health data to ad platforms on unfairness grounds, most recently in its complaint against Hims and Hers, and the theory there does not depend on whether the data was technically PHI.

This is where the usual clinic analytics stack falls apart. Google does not sign a BAA for Google Analytics 4, and any AI feature inside GA4 inherits that status. Consumer ChatGPT and consumer Claude.ai are not covered by a BAA either, so the habit of pasting an exported report into a chat tool for a summary moves treatment page data into a product that may use conversations for training unless the user opts out. If you want the detailed reasoning on running ads without a Google BAA, our piece on how hormone clinics can still run ads safely without a Google BAA applies almost unchanged to peptides.

We built Curve to sit on the other side of that line. Every customer on every plan gets a BAA. Events land server-side in US infrastructure before anything is forwarded, and what leaves is stripped down to what an ad platform is permitted to receive. Curve Analyst runs on Claude through Amazon Bedrock, inside infrastructure covered by Curve's BAA with AWS, and your data is not exported to a separate AI product or used to train models. It does not look up individuals or answer person-level questions, so your team can ask which treatment page converted best without anyone being able to ask who visited it. And we made it read-only, so a plain-language question cannot change your tracking or destinations by accident.

What to ask Curve Analyst

Open Analyst from the report you are already looking at and it keeps the date range and filters of that screen. Ask the weekly questions the way you would ask a colleague, and it answers with text plus a chart: a table for ranked lists, a step chart for funnels, a stat card for a total. Every number is a query against your account, it never estimates, and an empty metric is reported as empty. A fuller description of what it can and cannot do is in what Curve Analyst is, and the launch details are in the Curve Analyst announcement.

• Show me consult requests by traffic source last week, including organic, referral, and email.
• Which treatment pages had the most sessions and the highest consult rate?
• Which funnel step between treatment page and consult booked lost the most people?

Frequently asked questions

Can Curve Analyst tell me which of my peptide ads were disapproved?

No. Disapprovals and restrictions live inside the ad platforms, and Analyst does not read policy status. It shows spend, clicks, and Curve-attributed conversions by campaign, and a campaign whose spend goes to nothing partway through the week is the tell that something was pulled. Confirm the cause in the platform.

How do I see consults that came from ChatGPT or Perplexity?

Ask by source rather than by channel. Curve records the referring hostname, so chatgpt.com and perplexity.ai appear as their own rows in the sources report and Analyst can filter to them. At the channel level they roll into referral for now, because there is no dedicated AI assistant channel yet.

Will the consult numbers match what Google Ads or Meta reports?

They will not, and they are not supposed to. Curve reports its own attributed conversions using the model you chose, while a platform credits itself for anything within its attribution window. Analyst tells you which figure it is quoting, and the reconciliation view shows both together so you can watch the gap close as the window ages.

Can I find out whether a specific patient came from an email?

No, and that is a deliberate limit. Analyst answers aggregate questions and does not look up individuals. You can ask how many consult requests email produced and which campaign UTM drove them, which is the version of that question a marketing team needs.

Is any of this useful if we are barely running paid right now?

Yes, and arguably more so. Website analytics, goals, and funnels work with no ad spend at all, and for a clinic living on organic, referral, and email they are the whole picture. Campaign reporting and reconciliation pick up the moment a paid campaign is approved, and keep working through the next disapproval.

If you would rather answer these questions on Monday than reconstruct them from a pile of tools, book a demo of Curve and bring your treatment page list.