Paid campaign strategy

Month-one Google Search plan: "ecommerce seo agency" family

Given: agency selling ecommerce SEO · Google Search · USD 3,000/month (≈ USD 98/day) · landing page https://addition-labs.com/services/seo · zero conversion history. Source: the Google Search playbook (intent ladder, account structure, match-type gates, negatives, bidding by volume, offline conversions, QS/landing page rules).

0. Four things I still need before this goes live

The request covers goal, budget, landing page and current state. These are not covered and each changes a setting below. Answer them and the plan is launch-ready; until then the marked lines are assumptions.

1. Geography. Which countries/states should see the ads? (Sets location targeting; the plan assumes a single English-speaking market.)
2. What counts as a conversion, and is it tracked? The landing page has three different actions: a one-field email form ("Send me AI & SEO Audit"), "Schedule a meeting", and "Contact our team". Which one is the primary conversion, and is Google Ads conversion tracking (or GA4 import) installed and tested on it? Nothing launches before a test conversion fires.
3. Target CPA / what a lead is worth. Average deal value and close rate from lead to client. Without it there is no ceiling to judge month-one CPL against. (Not needed for the month-one bid strategy, which runs without a target anyway.)
4. Buyer language. Any sales-call phrases or existing search-terms data for how prospects describe the problem ("Shopify store not ranking", "product pages not indexing", etc.). Keywords below are built from the query family you named plus what the landing page itself claims; they should be checked against real buyer language before launch.

1. Where month one sits on the intent ladder

Spend opens rung by rung, and each rung unlocks only after the one below proves it converts.

2. Account structure

Two campaigns, each with an independent budget (never a shared budget: brand is the cheapest, highest-converting traffic and would eat the non-brand budget you actually need data from).

Account
├── GOOG_Search_Brand_SEO_2026-09            ~USD 5/day ceiling (assumption: low brand search volume; raise only if it caps daily)
│   └── Ad group: Brand                      "addition labs", "addition labs seo", "addition labs agency"
│
└── GOOG_Search_NonBrand_EcomSEO_2026-09     ~USD 93/day (the rest of the USD 3,000)
    ├── Ad group 1: Ecommerce SEO agency     the core query family
    ├── Ad group 2: Ecommerce SEO services   service-noun variants
    └── Ad group 3: Shopify SEO agency       platform-named variants (the page names Shopify explicitly)

Why one non-brand campaign, not three: the consolidation rule. A campaign that cannot reach roughly 15–30 conversions a month cannot feed smart bidding, so on a USD 3,000 budget with no history, the query family stays in one campaign with themed ad groups. Ad groups are themed (5–15 closely related keywords, one intent, one promise), not single-keyword ad groups. Only split ad group 3 into its own campaign later if it needs a different landing page or value proposition.

Campaign settings to flip on both campaigns at creation:

• Search Partners: off. Display Network: off.
• Location option: "Presence" (people physically in the target geo), not the default "presence or interest".
• Language: English (targets the user's Google interface language, not the query language; if you sell into non-English markets, that is a separate decision).
• Ad rotation: optimize. Two to three RSAs per ad group (see section 7 on generating them).
• Remarketing lists: build the audience now (all visitors to /services/seo, 30 and 90 days) so it has data by month two; do not run a remarketing campaign in month one.

3. Keywords and match types

Month-one rule: Phrase + Exact only. No Broad. Broad match is unlocked only when all three are true: 30+ conversions/month in the campaign, smart bidding live, and a tight negative list. An account with zero conversion history meets none of them.

Every keyword is added in both phrase and exact. Exact still matches close variants; phrase is the workhorse that catches the long tail while staying inside the intent.

Ad group 1: Ecommerce SEO agency ecommerce seo agency · e-commerce seo agency · ecommerce seo company · ecommerce seo firm · seo agency for ecommerce · ecommerce seo experts · ecommerce seo consultant · online store seo agency

Ad group 2: Ecommerce SEO services ecommerce seo services · e-commerce seo services · seo services for ecommerce · ecommerce seo service · seo for online store · ecommerce seo specialist

Ad group 3: Shopify SEO agency shopify seo agency · shopify seo services · shopify seo expert · shopify seo company · seo agency for shopify · shopify seo consultant

Brand ad group (exact + phrase): addition labs · addition labs seo · addition labs agency

Notes on the list:

• No search volumes or CPCs are quoted here; none were pulled. Check volumes in Keyword Planner before launch and drop anything with effectively zero volume rather than keeping it for completeness. A 50-search/month term with clear intent beats a 5,000-search term with mixed intent.
• Tag each keyword by intent tier in the label field so the search-terms ritual is faster.
• Do not put the same keyword at the same match type in two ad groups. Ad group 3 owns everything containing "shopify"; ad groups 1 and 2 get shopify as an ad-group-level negative so the query routes to one home. Same pattern for "services" vs "agency" if the search-terms report shows them colliding.
• Keywords not on the list that convert will appear in the search-terms report; that is how the list grows (section 5), not by guessing at launch.

4. Negatives

Two layers: a build-time starter list (below), and the query-specific negatives that come only from the weekly search-terms report. Nothing query-specific is invented here; there is no search-terms data yet.

A. Account-level negative list "Universal junk" (apply to the non-brand campaign; brand campaign gets the same list minus login): free · cheap · jobs · job · salary · hiring · career · careers · intern · internship · student · course · courses · tutorial · training · certification · pdf · template · reddit · wiki · login

B. Account-level negative list "Research intent": "what is" · "how to" · examples · meaning · definition · checklist · tips · guide · "best practices"

C. Campaign-level "Category collisions" for the non-brand campaign (terms the query family shares with things you do not sell):

• software · tool · tools · plugin · app · extension (people looking for an SEO tool, not an agency)
• diy · yourself
• freelancer · freelance (assumption: you do not want to compete for freelancer-rate queries; remove if you do)
• amazon · etsy · ebay (marketplace listing optimization is a different service; only keep this if you genuinely do not serve those, otherwise it is over-negativing)

D. Brand as a negative in the non-brand campaign: negative phrase addition labs so brand traffic is routed to the brand campaign.

Match-type mechanics to get right when adding these:

• Negative broad requires all its words to be present, in any order. Negative broad free trial does not block "free". Add single-word junk terms as single-word negatives.
• Negative phrase blocks the in-order phrase ("what is" as phrase blocks "what is ecommerce seo").
• Negative exact blocks only that exact query.

Do not over-negative. Every negative narrows reach, and at B2B volumes that compounds. Negative the clearly wrong, not the merely uncertain. amazon and freelance above are the borderline ones; the rest are clearly wrong for an agency selling done-for-you ecommerce SEO.

5. The weekly search-terms ritual (starts day 7)

Once a week, three passes on the non-brand campaign:

1. Waste: search terms with 3+ clicks and zero conversions. Negative the irrelevant ones (not the relevant ones that simply have not converted yet).
2. Winners: converting search terms that are not keywords yet. Add as Exact/Phrase in the right ad group.
3. Drift: phrase matches pulling adjacent-but-wrong meanings (for example, "ecommerce seo agency" pulling "seo agency pricing comparison" style queries). Tighten the match type or negative the drift.

Month one is where this ritual does most of its work, because the starter negatives above are generic and the real waste pattern is unknown until traffic runs.

6. Bidding for month one

Bidding is set by conversion volume, and the campaign is at zero.

Recommendation for weeks 1–2: Manual CPC with a per-keyword max CPC you set from Keyword Planner's top-of-page bid range for the target geo (not quoted here; pull it in the account). Reason: with no conversion history and unconfirmed tracking, Maximize Conversions has nothing to learn from, and a manual cap is the only thing that keeps the daily budget from being spent on a handful of expensive clicks. Switch to Maximize Conversions (no target) once the primary conversion fires reliably and the campaign is on track for 15+ conversions in the month. Do not set a Target CPA in month one; a target with no history either chokes delivery or is meaningless.

Rules of thumb that apply from day one:

• Smart bidding needs roughly 30 conversions in 30 days per campaign to learn. USD 3,000 may or may not get there in month one; the month-one job is to find out.
• Campaigns can spend up to 2× daily budget on a single day. Google balances to the monthly total; do not react to a single-day overspend.
• When you do move bids or targets, move in ±10–15% steps and wait 1–2 weeks. Every change restarts learning.
• Budget changes: ~20% at a time, never 30%+ in one move, 3–5 days between increases.

7. Conversion tracking and offline conversions (the highest-impact setup item)

• Primary conversion: pick one (see question 2). Given the page, the sensible primary is the meeting booking or contact form, with the one-field audit request as a secondary conversion. A one-field email form is a short form, and short forms buy volume at lower quality; if the audit request is made primary, smart bidding will optimize to email addresses, not pipeline.
• Test a real conversion before launch. This is the first item on the universal pre-launch checklist and the first item on the common-mistakes list.
• Capture GCLID on every form submission from day one and store it on the lead in your CRM. Month one will not have offline data to import, but the plumbing must exist now so that when leads turn into qualified opportunities or clients in months two to six (B2B clicks close in 60–180 days), you can import SQL/opportunity/closed-won back into Google with deal values. Until that loop exists, whatever bidding strategy you run optimizes to form-fills.
• Reconcile Google's conversion count against the CRM monthly. The CRM wins.

8. Landing page: what the fetch showed

I fetched https://addition-labs.com/services/seo (HTTP 200). Observations against the playbook's landing page rules:

• Title tag is "Ecommerce SEO Agency | Addition" and the meta description leads with "Ecommerce SEO that brings buyers, not just traffic … for Shopify stores". Both are on-query.
• H1 reads "Search Engine Optimization / Pages people want to read, that search engines can read too." That does not echo the query or the ad's promise. Message match is the QS lever that moves landing page experience: the H1 should carry "ecommerce SEO agency" (or the winning RSA headline once you have one), not a generic SEO line. The page already says "ecommerce seo agency" three times further down; it is the headline that is off.
• One job, one CTA: the page currently offers three distinct actions (email audit request, schedule a meeting, contact the team) plus a service-menu of eleven other SEO offerings (B2B SEO, SaaS SEO, Manufacturing SEO, Site Migration, and so on). For paid traffic, a page with one job and one CTA converts better; consider a paid-only variant of the page with the menu collapsed and a single primary action.
• Proof above the fold: case studies and "Our partners turn searches into visits, and visits into sales" exist as H2s; whether proof is visible above the fold I did not verify (no render, HTML only).
• Speed and mobile: not measured in this run. Both are pre-launch checklist items.

9. RSAs

Not generated here because you did not ask for them and the RSA output spec is a separate mandatory format (character limits, sidecar artifacts, self-check). When you want them: two to three RSAs per ad group, headlines built from the keywords in that ad group plus the page's own claims ("brings buyers, not just traffic", "Shopify stores", "run for you"). Ask and I will produce them to spec.

10. Month-one scorecard and milestones

Track exactly eight numbers weekly: spend · leads · CPL · lead→SQL rate (from CRM) · SQLs · cost per SQL · Search impression share · top wasted search terms.

Diagnostic to watch from week two: Search Lost IS (budget) vs Lost IS (rank). Lost to budget means the USD 3,000 is the constraint; lost to rank means Ad Rank (bid × quality) is, which you fix by fixing QS components before raising bids.

No benchmark numbers are quoted for CPC or CPL because the playbook's ranges are B2B SaaS practitioner starting points, not ecommerce-SEO-agency data, and the instruction is to anchor to your own first 30 days. Month one is that anchor.