Long-form answers to every question on a free-audit call.

Day one we get manager-level access to your Google Ads + GMC + GA4 + server-side tagging stack and run the deep audit on the live account. Day two through five we build the rebuild plan: campaign restructure, asset production schedule, AI ops setup, POAS targets, and the cross-channel readout cadence. By Friday of week one you have the plan in writing, and the first asset group rebuild or feed surgery is shipped or scheduled. Most agencies spend the first month on "setup" and onboarding theatre. We start moving spend by day five.

No, but you'll know within the first 30 days whether you want to. The free audit you receive before any engagement is structured to be diff-able against what your current agency is doing, same metrics, same deliverable, same time horizon. Most founders use the audit to compare directly. We don't ask you to commit to firing anyone. The numbers do the asking.

Manager-level Google Ads access, read-only GMC, GA4 read-only, and Shopify reports access if you're on Shopify. The audit is read-only, we don't change anything during the audit phase. If we move to the engagement, we ask for write access on Google Ads + GMC + GTM + the relevant CRM (Klaviyo, Postscript) and we sign your NDA. We'll never ask for raw login credentials.

Yes. Most ecom founders are wary of handing over account-level visibility before legal protection is in place, and we'd be wary too. Send your standard mutual NDA and we'll sign it the same day. If you don't have one we have a one-pager template that takes 5 minutes to review. The audit doesn't start until both sides are protected.

Yes, and many of our best engagements look exactly like this. The pattern that works: your in-house buyer keeps tactical day-to-day ownership of the account, Ad-Lab takes ownership of the system around the ad (landing pages, creative, feed engineering, server-side tracking, AI ops, weekly audits) and the strategic direction. The buyer gets a senior partner instead of being the only person in the account. We co-author the weekly readout. It works because nobody loses ownership of anything they were doing well.

Google Ads (Search, Shopping, PMax, AI Max, Demand Gen) and YouTube Ads as the paid acquisition core. Around that we run landing pages, video and image creative, feed and GMC engineering, server-side tracking, and the AI ops layer. We do not run Meta, TikTok, or LinkedIn ads, that's a hyper-focus discipline, not a limitation. We have specialist partners we trust on those channels and refer freely when a client needs them. The Google + YouTube work compounds whatever else you run; we report the cross-channel context every week.

Performance pricing means a base management fee plus a performance share tied to a POAS or revenue benchmark agreed at the start of the engagement. The exact split varies by spend tier and engagement scope and gets quoted on the audit call once we've seen the account. There are no retainers, no setup fees, and no charges for the system around the ad, landing pages, creative, feed work, tracking, those are included in the engagement fee.

Month two is when most accounts dip. The deep account rebuild we ship in week one usually breaks the prior agency's branded-search reporter pattern, which means the dashboard ROAS number drops while the actual incremental revenue rises. We forecast this and put it in writing on the audit call so you're not surprised. The metric to watch in month two is POAS and blended MER, not channel ROAS in isolation. Month three is when both numbers land, and they land higher than where you started.

$30,000+ in monthly Google + YouTube ad spend, or $100,000+ in monthly revenue. Below that, the system around the ad doesn't compound enough to justify our model, you're better served by a freelancer who can give you 20 hours a week of attention at lower cost. Above $1M/month is where the asset production layer and AI ops compound hardest, but the engagement structure stays the same regardless of spend tier.

Realistic ranges from the audit data we've seen across 350+ ecom accounts: 20-50% revenue lift in the first 90 days against the trailing baseline, with most accounts landing 30-40%. That comes from three places, fixing branded-search bleed (reveals the real non-brand opportunity), shipping landing pages purpose-built for the click intent, and surfacing creative angles that the prior account never tested. If your account is already in the top decile of optimisation, the lift is smaller. If it has the leaks we typically find, the lift is larger.

No. Landing page production (5-10 per client per month), video creative (18 angles per batch), image generation across PMax and Shopping, and feed engineering are all included in the engagement fee. We don't subcontract any of it, and we don't upcharge any of it. The reason: the work that compounds Google + YouTube spend at scale is the system around the ad, and charging separately for the system creates the wrong incentive. We win when the ad spend earns more, not when we sell more sub-services.

GMC suspensions happen for a small set of well-known reasons (misrepresentation, policy violation, store quality, item-level pattern flags) and there's a recovery playbook for each. We diagnose the root cause from the suspension notice + feed audit, surgical-fix the feed at GMC level (not your Shopify export), and write the reconsideration request, Google reviewers respond to specific evidence, not boilerplate appeals. Recovery typically takes 2-7 days. Account-level suspensions on accounts we've onboarded are tracked daily; we don't let suspensions sit. If you come to us with an active suspension we can usually start the recovery work the same day.

Yes, that's the modal stack of our portfolio. Shopify Plus for storefront, Klaviyo for email + SMS, Northbeam (or Triple Whale, or Polar) for attribution, GA4 for behaviour, GTM Server-side for first-party tracking. We integrate at the data layer, not the dashboard layer: we feed Northbeam-attributed conversions back into Smart Bidding via Conversions API where the math improves Smart Bidding's learning, and we leave the dashboards alone otherwise. The mistake most agencies make is trying to optimise against three attribution sources at once. We pick one for bidding decisions and use the others for board reporting.

Yes. GTM Server-side migration is part of standard onboarding for any account that's still browser-side in 2026, included in the engagement fee. Implementation runs 1-2 weeks: provision the GTM Server-side container, wire Conversions API for Google + Meta + TikTok where applicable, configure Enhanced Conversions, set up customer-list sync, validate EMQ scores against benchmarks. We do the engineering; your team approves the configuration. Once it's live, every Smart Bidding decision and every attribution number gets cleaner, most accounts see a 5-15% recovery in attributed conversions just from the cutover.

Consent Mode v2 is a hard requirement for EU traffic since March 2024. We configure the consent signal to flow from your cookie banner through GTM into Google's tags, validate it in DebugView, and verify the consent_mode parameter is showing on incoming hits. Most accounts we onboard have it half-configured, banner shows, signal doesn't reach Google. The fix is engineering, not legal. Once it's working, EU traffic counts in your dashboards and Smart Bidding can learn from it. Until it's working, you're flying blind on a chunk of revenue and risking compliance.

The honest answer is geo-holdout testing, turn off ads in a matched market for 4-6 weeks and measure the revenue delta against the control. That's the only attribution method that survives every cookie change, model swap, and iOS update. We run holdout tests on a rolling basis for accounts above $100K/month, typically one channel at a time, with the test design and decision rules locked at the start. For lower spend tiers, we use brand-vs-non-brand splits, post-purchase attribution surveys, and the cross-channel MER trend as proxies. The proxies are noisier than holdouts but they earn their place when the volume isn't there for clean experiments.

Probably yes, but not before the structure work is in place. AI Max went GA April 15, 2026 and changes how match types and asset selection work inside Search and Shopping. Turn it on without the brand-exclusion list and feed labelling first and your blended ROAS drops in week one because AI Max cannibalises your existing PMax. Turn it on after the structure work and most accounts see 8-15% incremental conversion volume in 60 days. The configuration that matters: brand-exclusion list, feed-level labelling for AI Max-eligible inventory, audience-signal seeding, and a defined success metric before you flip the switch. We do this as a structured launch, not a checkbox.

Yes. Every engagement has a dedicated Slack channel staffed seven days a week and a weekly Loom walkthrough sent every Monday with the prior week's numbers, what we changed, and what changes next. The Loom is recorded by the buyer running your account, not a junior. The Slack channel is for ad-hoc questions; we respond same-day during the working week and same-week on weekends. We don't use a portal because nobody opens portals.

Day 30: the audit is done, the rebuild plan is in motion, the first landing pages and creative angles are live, server-side tracking is migrated. The dashboard ROAS often dips slightly because we've broken the prior branded-search reporter pattern. Day 60: blended MER is up, POAS is up, the asset group structure is settling into Smart Bidding's learning. Most accounts cross their trailing baseline this week. Day 90: full-stack performance is compounding, the AI ops layer is producing weekly reports without manual work, and we have enough data to formalise targets for the next quarter. The exact numbers vary by account; the shape of the curve is consistent.

Month-to-month. The 90-Day Account Rebuild is the structured kickoff; after that we run on a rolling monthly basis with no minimum commitment beyond the initial 90 days. Most clients stay 12-24 months because the system compounds and switching costs are real. Some leave at month 6 because they've internalised the system and want to run it in-house, which is a fine outcome. We don't penalise either path. The contract is clean: 30-day notice, no termination fee.

Stalls happen and they're usually caused by one of three things: market-level demand drop (we surface it on the cross-channel readout), creative fatigue (we ship the next batch of 18 angles), or the account hitting the structural ceiling for its current spend tier (we re-architect the asset groups). The diagnostic is in the weekly readout and the response time is days, not months. If a stall is genuinely structural, the brand has hit its TAM or the unit economics no longer support the current CAC, we say so and recommend the strategic move. We're not in the business of running spend that doesn't work just to keep the engagement.

Yes for the median, no for the outliers. The case studies on the homepage are real numbers from real clients, but they're not the average, they're the wins worth surfacing. The realistic expectation is closer to 30-40% revenue lift in 90 days, not the 10x outcomes you'll see on the case study cards. Those exist; they're not the median. We give you the median forecast on the audit call against your account specifically, with the upside and downside scenarios both written down.

Weekly Loom every Monday. Slack channel every working day for ad-hoc questions, same-day response. Monthly written readout with cross-channel context (POAS, ROAS, MER, AOV, what we changed, what changes next). Quarterly strategic review with the next quarter's targets and structural moves. Spend drops or conversion drops trigger same-day Slack alerts. We don't show up only when there's good news, and we don't disappear during slower months. The cadence is structural, not vibes-based.