;
};
Publishing a page is the start of a loop, not the end of one: traffic arrives, some of it converts, and the numbers tell you what to change next. This page covers where those numbers live, which signals to check after a launch, and how to have Replo run the report for you every week instead of pulling it by hand.
## Where your numbers live
Three places report on a Replo site, and they answer different questions:
* **Insights, inside Replo.** Replo tracks page views, sessions, add-to-carts, checkouts, and purchases on every site it publishes, automatically, with nothing to install. The [Insights app](/apps/insights) turns that into dashboards: revenue, conversion rate, sessions, and average order value up top, plus reports covering top landing pages, referrers, and UTM sources. This is the fastest answer to "what happened on my Replo pages this week."
* **Your own analytics tools.** GA4, the Meta Pixel, and attribution tools like Triple Whale keep working on Replo pages once you add your existing IDs in **Scripts & Pixels** ([Adding Scripts](/features/scripts)). Because Replo pages live on a subdomain of your own domain, the same cookies and visitor identity carry across your store, your Replo pages, and checkout; [Why Publish to a Subdomain](/why-subdomains) walks through the mechanics.
* **Shopify Analytics.** Orders from Replo pages are real Shopify orders, so revenue, orders, and conversion attribution show up in Shopify as usual. Page views on Replo pages do not, because Shopify only reports on pages it hosts. Use Insights or GA4 for traffic, and Shopify for order-level reporting.
## Before you begin
1. [Publish your site](/publishing). Replo's own tracking starts on publish.
2. Add your pixels in **Scripts & Pixels**: the same Meta, GA4, and attribution-tool IDs you use on your main store ([Integrations](/integrations/introduction)).
3. Keep UTMs on your ad URLs exactly as you do today. Replo forwards UTMs and click IDs through to checkout, so attribution survives the handoff ([details](/why-subdomains)).
4. Optional but worth it: connect your analytics tools in the [Integrations app](/apps/integrations) so Replo can read their data in chat, separate from tracking with them. GA4, Triple Whale, Northbeam, PostHog, Amplitude, Klaviyo, Microsoft Clarity, and Google Search Console are all readable once connected ([GA4 setup](/integrations/ga4)); the section below shows why that's worth doing.
## What to check after a launch
Each signal below has a place to look and an action to take. All of them work as prompts too; Replo builds the report and saves it in Insights so you can re-slice it later.
### Which pages get traffic
Open the Overview report in [Insights](/apps/insights) and look at your top landing pages. Two failure modes look similar from the outside and need opposite fixes:
* A campaign page with almost no sessions has an ad problem, not a page problem. Check the ad's destination URL and whether the campaign is actually spending.
* A page with sessions but little revenue is where page work pays off. The next two signals narrow down why.
```text theme={null}
Show which landing pages are driving revenue this week, broken down by traffic source.
```
### Where visitors drop off
Replo tracks the on-page funnel (page view, add to cart, checkout started, purchase), so you can see where a page loses people instead of guessing:
```text theme={null}
Build a report showing the funnel from page view to add to cart to checkout to purchase for my top landing pages over the last 14 days.
```
Reading the result: visitors who leave without adding to cart point at the page itself, so look at the offer, the first screen, and [load speed](/features/page-speed). Visitors who add to cart but never start checkout point at the cart, so check its pricing, shipping messaging, and upsells. The Sessions report's average duration and bounce rate tell you whether people are engaging with the page at all before they leave.
### Which traffic sources convert
The Overview report breaks sessions down by referrer and by UTM source, and conversion by campaign is one prompt away:
```text theme={null}
Which UTM campaigns drove the most purchases last month?
```
Move spend toward what converts, and treat a high-traffic zero-conversion source with suspicion: it is usually a mismatch between the ad's promise and the page it lands on. If you connected GA4 read access, cross-check against Google's numbers by asking Replo to build a report using your @Google Analytics 4 connection.
## Each connection makes every report smarter
Every analytics tool you run answers one kind of question. Insights knows your on-page funnel. GA4 knows your traffic. Triple Whale or Northbeam knows attribution. Klaviyo knows your email list. Normally those answers live in separate tabs, and joining them is a spreadsheet job that never happens.
Connect them in the [Integrations app](/apps/integrations) and they stop being separate tabs: Replo reads each one in the same chat that reads Insights, so one prompt can pull several tools into one answer.
* **Cross-check the numbers you're about to act on.**
```text theme={null}
Compare last month's revenue and conversion rate in Replo analytics
against GA4 and Triple Whale, and flag anywhere they disagree.
```
* **Follow one campaign across tools.** Klaviyo knows which emails went out; Replo knows what each UTM earned.
```text theme={null}
Pull my Klaviyo campaigns from last month and match them against landing
page revenue by UTM. Which campaigns are worth repeating?
```
* **Widen the lens when a page underperforms.** Search Console shows what people searched before they landed; Clarity, PostHog, or Amplitude show what they did on the page.
```text theme={null}
This page gets traffic but doesn't convert. Check Search Console for the
queries sending people here and tell me if the page matches the intent.
```
This is where connections compound: each one you add makes every report after it smarter, because any prompt can now draw on it. Two tools connected means cross-checks; four means one weekly report that would otherwise take an analyst an afternoon.
## Put the reporting on a schedule
You do not have to pull these reports by hand. A [task](/apps/tasks) is a prompt with a schedule; when it fires, Replo runs it in a normal session with the same access it has in chat, including your analytics. So a weekly performance summary is one prompt:
```text theme={null}
Every Monday at 9am, summarize last week's traffic, revenue, and best-converting pages, and flag anything that dropped.
```
Tasks repeat daily, every weekday, weekly, or monthly. Each run's report lands in the task's history: open the run in Tasks and read it like any chat session. And if the report should reach your team, not just you, connect [Slack](/integrations/slack) and end the prompt with "post it to our #marketing channel": the report shows up where your team already works.
A run is a full session, so a task can also do the work instead of reporting on it:
```text theme={null}
Every Friday, check my published pages for broken links and slow images, and report what you find.
```
For the full loop (ship pages, watch Insights, let scheduled reports flag what changed), see [Sell at Scale](/use-cases/sell-at-scale).
## Next steps
Dashboards for traffic, revenue, and what converts.
Schedule prompts to run in the background.
How tracking, ads, and checkout keep working.