The infrastructure
behind every account we run.
ADSRUNNER is an agency, but the work behind your account runs on a platform we built in-house. It reads your ad accounts, your analytics, your store, our first-party tracking, and your customer list side by side, checks them every day, and ranks what it finds by the money at stake. A senior strategist decides what changes. It is running on client accounts today.
Most agencies are running 2018 operations on 2026 platforms.
The platforms changed. Performance Max made campaign control opaque. iOS 14 broke pixel attribution. Smart Bidding shifted control from media buyers to algorithms. AI compressed what used to take days into seconds.
But most agencies still operate the way they did before all of this. Spreadsheet-based account management. Weekly manual reviews. Reporting decks built once a month. The work that used to make sense in 2018 is now structurally too slow for how the platforms operate.
We built our platform because we needed it to do our job well. Not as a SaaS product to sell — as the infrastructure that runs underneath every client engagement we take on.
Four ledgers, read side by side, never averaged.
Every account has at least four versions of the same month. Most reporting picks one, or blends them into a number nobody can audit. We keep them apart, in one view, on one set of definitions.
The disagreement between them is the finding. It shows where a channel's claim is in the right neighborhood, where the tracking has broken, and where the business is paying for sales it would have made anyway.
- 01
What the platform claims
Google, Meta, Microsoft, TikTok, and Pinterest each report the sales they believe they caused, on their own attribution windows. Useful, and graded by the party being graded.
- 02
What analytics recorded
GA4 sees sessions, sources, and landing pages across every channel, with its own rules for credit and its own delays.
- 03
What our tracking saw
First-party tracking on your domain, with purchases also captured on the server, so a browser setting does not quietly delete a sale.
- 04
What the store banked
Orders, line items, product cost, refunds, and net revenue from Shopify. The number the business is judged on.
The method in long form: why Meta, Google, GA4, and Shopify report different revenue.
Five capabilities, one connected system.
Every part feeds the next. Tracking and connectors fill the data layer. The daily checks read the data layer. What they find becomes a proposal. A person decides, and the client sees the same numbers we do. None of it works in isolation, and none of it spends money on its own.
Daily checks, ranked by dollars
Twenty-seven automated checks read every account every day: spend pacing, wasted search terms and placements, creative fatigue, feed failures, disapproved ads, conversion drops, and tracking health, across Google, Meta, Microsoft, TikTok, and Pinterest. Each finding is compared with the account's own history and ranked by the money involved. A large swing on a tiny campaign stays small.
First-party tracking and attribution
Tracking on your own domain records visits and purchases in data you own, with purchases also captured on the server so a browser setting does not quietly delete them. Six attribution models run on the same data, side by side: last click, first click, linear, time decay, position based, and data driven.
One data layer, one set of definitions
Spend and results from five ad platforms, sessions from GA4, and orders, line items, product cost, refunds, and net revenue from Shopify land in one analytical store with the same columns. A campaign from Google and a campaign from Meta can finally be compared on the same terms, next to what the store banked.
Your customer list, matched
Connect your customer list and every purchase is classified as a new customer, a returning one, or one we cannot yet tell apart. That turns cost per new customer from a guess into a number, and shows whether extra budget is buying reach or renting loyalty. A pixel alone cannot know who bought from you last year.
A client portal instead of a monthly deck
Clients see the same account we see: performance, campaigns, creative, analytics, attribution, reports, and approvals, in one place. A read-only assistant answers plain-English questions from the data, with a hard rule against stating a revenue figure it has not checked. Excel and PDF reports can run on a schedule.
Sense. Propose. Approve. Ship.
These are the working surfaces of the platform: the activity feed our team watches, the approval card a strategist signs off on, and the portal a client sees. Every change follows this loop. The checks sense the signal, a proposal drafts the fix, and a senior operator approves it before anything ships.
Prospecting → High-intent converters
Prospecting → High-intent converters
You don't buy software. You buy a partnership that runs on it.
Our platform is the infrastructure underneath every managed engagement. Clients get the portal, the reports, and the assistant as part of the work. It is not a separate product to evaluate. It is the system that makes our team better at the work.
A calm window into your account
A portal refreshed every day across every connected platform. An approval queue for any change we propose. Scheduled reports, and an assistant that answers questions about your own account. Outcome-first language, revenue, customers, and growth, rather than platform jargon.
A craftsperson's workshop
Every client account in one cockpit. A ranked list of what changed overnight, so a strategist starts the day on the three things worth money instead of a blank login screen. The watching is automated. The judgment is not.
Owned, not rented
Your data is isolated to your account and encrypted. The tracking sits on your domain and the customer list is yours. We never sell, share, or aggregate client data with third parties. We have to be worth keeping, because we are not hard to leave.
Security and isolation, not optional.
We run client data the way a financial-services platform runs its own — encryption at rest, row-level isolation, an immutable audit trail, and AI that inherits the requesting user's permissions.
Multi-tenant isolation
Every database query scoped by account. No client data accessible across tenants. Row-level security enforced at the database layer.
Encrypted credentials
OAuth tokens and API credentials encrypted at rest with AES-256. Never exposed to frontend code. Server-side API calls only.
Full audit trail
Every proposal, every approval, and every campaign change is logged. Reviewable at any time by you and us.
Permission-aware AI
AI agents inherit user permissions. Cannot access data the requesting user is not authorized to see. Read-only by default for clients.
The questions most agency setups cannot answer.
These become askable once the ad accounts, the store, the website, our tracking, and the customer list sit in one picture and someone looks at it every day. Each answer is an input to a strategist, not a change to your account. The full list, with how to test your own agency against it, is in fourteen questions your agency should be able to answer.
- Was last month actually profitable once product cost and returns are in?
- Ad platforms report revenue with none of the costs underneath it, so a strong return on ad spend can still be a losing month. We pull orders, product cost, refunds, and net revenue from the store and set them next to spend, so the month is judged the way the business is judged.
- What did it really cost to win a new customer?
- Blended return on ad spend counts a loyal customer who was coming back anyway as an acquisition. With your customer list, hashed on the way in, we separate people who had bought before from people who had not, and report cost per new customer by source.
- Four platforms claim the same sale. Which one earned it?
- We record the journey on tracking you own and show six ways of giving credit side by side: last click, first click, linear, time decay, position based, and data driven. Where they disagree tells you how much of your confidence is earned. We do not treat any model as truth until it agrees with the store.
- Why do the ad platform, analytics, and the store disagree?
- They count different things, with different delays and different incentives. We show what the platform claims, what analytics recorded, what our tracking saw, and what the store banked in one view, instead of blending them into a figure nobody can audit.
- Did sales drop, or did the tracking break?
- A fall in purchases while page views continue is usually a measurement failure, not lost demand. Because the site and the store sit next to the ads, we can tell a broken signal from a real drop and fix the measurement before anyone rescues a campaign that was never sick.
- Where is money going out with nothing coming back today?
- Search terms with spend and no conversions, placements that never return, landing pages that convert nobody, and disapproved ads the budget still tries to spend. We look for them every day and rank them by dollars, not by how dramatic the percentage looks.
- Is growth coming from new demand, or from people who already buy?
- Once new and returning customers are split by channel, you can see whether extra budget is buying reach or renting loyalty. Retargeting often looks brilliant until this split exists.
- Does the platform change my ad accounts on its own?
- No. The ability to carry out changes automatically is built and deliberately switched off. The platform finds and ranks what matters, a senior strategist decides, and changes that affect you go through an approval you control. Every action is logged.
Live data sources today.
- Google AdsSearch, Shopping, Performance Max, YouTube
- Meta AdsCampaigns, ad sets, ads, breakdowns
- Microsoft AdvertisingSearch and Shopping, share metrics
- TikTok AdsCampaigns, ads, placements, video
- Pinterest AdsCampaigns and ads
- GA4Sessions, sources, landing pages, conversions
- ShopifyOrders, line items, product cost, refunds, net revenue
- First-party trackingOn your domain, with server-side purchase capture
- Customer listEmails and phones hashed before storage
LinkedIn advertising is work we run as a service rather than a data connection. How the store connection works today: Shopify on ADSRUNNER.
Shipped, not promised.
- Now
Shopify App Store listing in review
The public Shopify app is submitted for review. Until it is listed, stores connect through a read-only custom app the owner approves.
- Aug 2026
Daily checks expanded to 27
New checks for wasted placements, landing pages that stop converting, keyword efficiency, disapproved ads, and revenue sources that disagree with each other.
- Jul 2026
New versus returning customers
With a customer list connected, every purchase is classified as new, returning, or not yet identifiable, and cost per new customer is reported by source.
- Jun 2026
First-party tracking and six attribution models
Tracking on the client domain with server-side purchase capture, and six attribution models compared side by side on the same data.
- May 2026
TikTok, Microsoft, and GA4 in one layer
TikTok Ads, Microsoft Advertising, and GA4 joined Google, Meta, and Shopify in the same data layer, on the same definitions.
What we build next is decided by the questions clients ask that we cannot yet answer well. We do not publish dates for work that has not shipped.
Ready to see it in action?
Book a strategy call and we'll show you the platform live — against your actual account data, not a generic demo.