Forgemark

Features

One app, eight questions answered

Each screen answers one question with your own numbers, and every finding comes with a fix. These are real Forgemark screens, shown with our sample stores: Northfield Supply Co. for a Shopify shop, and Larchmere Kitchens for a business that wins enquiries.

01

How often AI assistants recommend you

Every week Forgemark asks ChatGPT, Gemini, Google's AI Overviews and AI Mode, and Perplexity the questions your shoppers ask, as a shopper in your country would. You can add Claude too.

You see your share of answers, the place you're named in, the rivals named instead, and the pages each engine cites. Forgemark finds your rivals for you, from the brands the answers keep naming.

Forgemark's AI visibility screen: share of answers down from 33% to 28%, the store's line against its rivals, and share of answers for each of six AI engines over 13 weeks. Forgemark's buyer questions table: each question a shopper might ask, with a marker for every engine showing whether the store was named and in what place.
AI visibility for Northfield Supply Co., a sample store: the headline share, each engine's trend, and every buyer question engine by engine.
03

What your ads made, counted five ways

Forgemark reads your spend from Google Ads and Meta and credits each order five ways: first click, last click, linear, time decay and position based. Each count sits beside what the platform claims for itself, so you can see when two platforms claim the same sale.

MER and new-customer ROAS sit at the top, and Klaviyo's claim sits beside the credit Forgemark gives email. A direct visit only gets the credit when nothing else in the journey could.

Forgemark's ads and attribution screen: blended MER by day, ad spend, sales, MER, new-customer ROAS, cost per new customer and profit after costs, and a channels table comparing Forgemark's count with what Meta and Google Ads report.
Ads and attribution for Northfield Supply Co., a sample store, with the linear model selected.
04

Every step before each order

For every order and every enquiry, Forgemark lists each visit that came before it: the ad, the search, the email or the AI assistant, in the order they happened.

You see how many visits a sale takes, how many days, and the paths people take most often. Open any order to see what happened in each visit and how each attribution model shares the credit.

Forgemark's journeys screen: 2.2 visits before the average purchase, orders with every visit recorded, days to buy, the share that crossed platforms, and the commonest paths from first visit to order. One order opened in Forgemark: six visits over 3.9 days, from Perplexity, a direct visit, Meta ads and Google search, each with the pages seen, and how each attribution model shares the credit.
Journeys for Northfield Supply Co., a sample store: the commonest paths, and one order opened with every visit and each model's share of the credit.
05

Where visits stall

A Shopify app pixel and theme app embed, on analytics consent only, follow each step from visit to purchase and show where more shoppers leave than usual, with the likely cause.

Rage clicks name the element people keep tapping. Page speed comes from real visitors' phones and computers, page type by page type.

Forgemark's shoppers screen: sessions from visit to purchase, the steps losing more shoppers than usual with the likely cause, rage clicks by element, and page speed by page type.
Shoppers for Northfield Supply Co., a sample store.
06

Every enquiry, and what brought it

For sites that win work rather than sell online. One line of code counts every form sent, booking made and phone number tapped, and the search, page or AI assistant that brought each one.

Forms show how many people start and send each one, and the field they were on when they left. Forgemark reads field names only, never what anyone types.

For enquiry sites
Forgemark's enquiries screen: 57 enquiries, the enquiry rate, forms and bookings, calls and what the work is worth, how people get in touch, and where enquiries come from, AI assistants included. Forgemark's forms block: for each enquiry form, how many people saw it, started it and sent it, and the field they were on when they left.
Enquiries for Larchmere Kitchens, a sample kitchen business in Leeds.
07

Weeks and months against the year before

Pick a metric and weeks or months, and Forgemark draws it across your whole history with the same period a year before dashed beside it. Sales, orders, new customers, clicks and AI visits all have one.

A month that's only half over is marked "so far", never compared as if it were whole. So a good month is measured against the right one.

08

The fixes worth doing first

Every screen feeds one list: the surest fixes first, each with what it's worth in sales a month, the evidence behind it, and how sure we are.

Make the change yourself, or send it to Forge and it comes back as a costed job.

Forgemark's recommendations screen: £8,700 to £21,000 a month across eight fixes, surest first, each with its value, confidence, evidence and the fix.
Recommendations for Northfield Supply Co., a sample store.

See what it connects to

Shopify, WordPress, WooCommerce, Astro or any other site, plus Google, Google Ads, Meta and Klaviyo. The full list is on the integrations page.