Part 1 of 6
The 4 Google Ads Foundations You Must Fix Before Touching a Campaign
Direct answer
Before building or optimizing any Google Ads campaign, four things have to be correct: conversion tracking (one primary, value-based purchase event), your Google Merchant Center feed (titles, custom labels), your website's baseline conversion rate, and your account's quality score inputs (expected CTR, ad relevance, landing page experience). Get any of these wrong and every campaign decision built on top of them is built on bad data.
This is part 1 of a 6-part series covering the entire framework — see the full playbook index — but this part matters most, because nothing downstream works if this is broken.
Foundation 1: Conversion tracking
This is the single most important thing in your account. If your tracking is wrong, every optimization decision — every bid, every budget shift, every "this campaign is winning" call — is based on lies.
Common breakages I find in audits:
- Double-counted conversions (the thank-you page loads twice, firing the purchase event twice)
- Add-to-cart events tagged and optimized as purchases
- Google Analytics goals imported with session-based attribution instead of transaction-based
- Reliance on Google's "auto-detected" conversions instead of a custom, verified implementation
The correct setup: one primary conversion action, tagged as "Primary" in Google Ads: a completed purchase, with dynamic value passed (not a static value). Everything else — add to cart, begin checkout, email signup — gets tagged "Secondary." Secondary actions still provide signal to Smart Bidding, but they don't inflate the metric your account is actually optimizing toward.
Verify it with Google Tag Assistant before you touch a single campaign. Fire a test purchase. If the tag fires once, correctly, with the right value: you're clear. If not, fix this before spending another dollar on anything else in this series.
One layer deeper than the original checklist: as of 2026, Google's Smart Bidding models increasingly lean on Enhanced Conversions and value-based signals, not just raw conversion counts. If you're planning to use broad match or automated bidding anywhere in your account (see Part 4 on the bidding ladder), Enhanced Conversions isn't optional anymore: it's a prerequisite for the algorithm to have anything reliable to model against.
Foundation 2: your Google Merchant Center feed
For shopping campaigns and Performance Max — where most ecommerce revenue lives — your product feed is the campaign. The titles, descriptions, attributes and images in GMC determine which search queries you appear for, how often, and how relevant you look to a buyer mid-search.
Title optimization case study
Moving non-branded keywords into the front of every product title took one client from 9.63x to 18.97x ROAS. The logic: Google weights roughly the first 60-70 characters most heavily for query matching (this is the current, more precise figure — earlier guidance said "first 50 characters," which undersells how much room you actually have), and mobile shopping results often truncate around 4-5 words. If your brand name is eating those first words, you're telling Google to match you against branded queries instead of the category and problem queries where new-customer volume actually lives.
The formula:
[primary converting keyword] for [benefit] - [differentiator] - [specs] - [brand, if room remains]Example: "magnesium supplement 120 capsules — Brand X" becomes "magnesium bisglycinate for deep sleep and anxiety — 400mg — third-party tested — 90-day supply."
One nuance worth adding: if your brand is already well-known to your target buyer, leading with the brand name in title 1 can still outperform — branded search intent converts on recognition, not keyword match. Lead with product type and benefit for cold/category traffic; lead with brand only where brand recognition is doing real work (retargeting feeds, branded campaigns).
Custom labels are the other feed element most brands ignore entirely. Add labels for:
- Margin tier (high / medium / low)
- Bestseller status (yes / no)
- Use case (sleep / energy / recovery, etc.)
- Price tier (budget / mid / premium)
These labels are what let you run tiered ROAS targets by product: your highest-margin bestsellers get the most aggressive bidding, your low-margin catalog doesn't drain budget it can't earn back. Part 2 covers exactly how to structure this.
Test safely: use a supplemental feed in GMC to trial title changes without touching your main product data source. Build a Google Sheet with product_id and title columns, upload it as a supplemental feed, and it overwrites your main feed's titles: fully reversible if a change underperforms.
2026 update
Performance Max changes how this feed work pays off. Standard Shopping campaigns are still available and still the right call when you want full manual control over product groups, negatives and bids at the SKU level. But Performance Max now runs at a structural priority above Standard Shopping in the auction, and most net-new ecommerce spend is flowing through it. PMax doesn't give you Standard Shopping's manual listing-group bidding: instead, it reads your custom labels through asset groups and conversion-value optimization. Practically: your feed hygiene work (titles, labels, custom labels) matters more under PMax, not less, because it's the only lever you have left to signal product priority to the algorithm. Build your custom label taxonomy regardless of which campaign type you run — it's the connective tissue between this foundation and Part 2's tiering strategy.
Foundation 3: your website has to convert
Ads are a delivery truck. You can't blame the truck if the product inside doesn't sell itself, or if nobody can get the box open.
Before you scale any campaign, answer honestly:
- What's your conversion rate on direct/organic traffic? (This is your baseline: paid traffic on a relevant landing page should approximate it.)
- What's your mobile page load time? (Above 3 seconds is a real conversion killer.)
- Is your primary CTA above the fold on mobile?
- Are reviews visible without scrolling on mobile?
If your direct-traffic conversion rate is below 2% for an ecommerce store, your "Google Ads problem" is actually a site problem. Spend on fixing the site before you spend more on ads: no campaign architecture fixes a leaking bucket.
Foundation 4: account-level quality score
Quality score is the lever most brands never actively manage, even though it directly sets your effective CPC (higher QS = lower CPC for the same ad rank) and your eligibility for top-of-page placement.
Three components, three fixes:
Expected CTR → headlines that mirror the search query exactly. If someone searches "magnesium bisglycinate for sleep," Headline 1 should say "magnesium bisglycinate for sleep" — not a clever variation of it.
Ad relevance → single-theme ad groups (STAGs) instead of broad ad groups stuffed with dozens of loosely related keywords. Each ad group should hold keywords with closely related intent, and the copy should speak directly to that intent.
Landing page experience → match your landing page headline to your ad headline. If the ad promised "magnesium for 2am wake-ups" and the landing page says "buy our premium supplement," that gap costs you quality score — and conversions. Match the message, word for word if you can.
FAQ
What's the most common Google Ads mistake ecommerce brands make?
Building campaign strategy on top of broken conversion tracking. If the primary conversion event is double-firing, tracking add-to-carts as purchases, or using session-based attribution, every subsequent optimization decision — bids, budgets, "winning" campaigns — is based on inaccurate data.
Do I need Performance Max or Standard Shopping in 2026?
Both still work. Standard Shopping gives you full manual control over product groups, negatives, and bids by SKU. Performance Max now runs at a higher auction priority and captures most net-new ecommerce ad spend, but controls tiered bidding through custom labels and conversion-value optimization rather than manual listing groups. Many mature accounts run both, structured to avoid product overlap.
How many characters should a Google Shopping product title be?
Up to 150 characters is allowed, but the first 60-70 characters carry the most weight for query matching and are what mobile shoppers actually see before truncation. Front-load your highest-intent, non-branded keyword there.
What conversion rate should an ecommerce site have on direct traffic?
There's no universal number, but under 2% on direct/organic traffic is a signal that the site — not the ad account — is the constraint. Fix the site before scaling ad spend.