How do I restructure a failing campaign?

Part 1: Diagnose why the campaign is failing (before you “rebuild” anything)

Start with measurement and goals, because a “failing campaign” is often a tracking problem

If I inherit a campaign that looks dead on paper, my first question isn’t “Should we reorganize?” It’s “Are we measuring the right thing, the right way?” If your conversion setup is incomplete, double-counting, or sending low-quality actions into the main Conversions column, every decision you make afterward (keywords, ads, landing pages, bidding) gets distorted.

Make sure you’re optimizing to the actions that truly represent business value. If you have multiple valuable actions (for example, purchases and leads), you’ll usually get more stable optimization when each action has an appropriate value assigned and you use value-based bidding where it makes sense, instead of forcing a campaign to optimize to a mixed set of “upper and lower funnel” actions with no values.

• Confirm what’s “Primary” vs “Secondary” so only the right actions guide bidding and show in the main Conversions column.
• Verify the conversion source (website, calls, app, offline) and that it matches how customers actually convert.
• Check for tagging gaps (missing parameters, inconsistent checkout flows, duplicate firing, or recent site changes that broke tracking). If you’re using cart-data-style conversion details, validate that key parameters (item IDs, prices, quantities) are consistently passed.
• Validate conversion delay (the time from click to conversion). This matters for reading performance trends and for any Smart Bidding troubleshooting.

Check for delivery blockers that make performance look worse than it is

Before you restructure, confirm the campaign can actually serve consistently. Policy issues, disapproved ads, or “limited” statuses can quietly throttle delivery, especially if your strongest ads are the ones impacted. Use Policy Manager to identify what’s restricted and appeal only when you’ve either corrected the issue or you’re confident the decision is wrong.

Also look for basic “plumbing” constraints: budgets that are too tight for the objective, targeting that’s overly narrow (locations, schedules, audiences), or bidding targets that are set unrealistically compared to what the account can currently achieve.

Use search intent evidence, not guesses: the search terms report is your truth serum

If you’re running Search campaigns, the fastest way to understand why you’re bleeding money (or getting no traction) is the search terms report. Two important nuances matter here. First, broader match types can match search terms in narrower ways, so the “match type” you see in reporting is about the search term’s relationship to the keyword that triggered, not necessarily the keyword’s configured match type. Second, modern match behavior is meaning-driven: phrase match can include searches that match the meaning of your keyword, and broad can expand even further—so you must manage intent with structure and negatives, not nostalgia for “exact means exact.”

I’m looking for three patterns: wasted spend on irrelevant intent, spend clustering on a few themes that deserve their own structure, and “good intent” searches being under-served because ad relevance or landing pages don’t align.

Part 2: Rebuild the structure around intent, value, and controllability (not just “neater ad groups”)

Pick a structure model that matches how you want to control budgets and performance

A restructure should give you clearer levers: budget control, query control, and message-to-landing-page alignment. In practice, that usually means separating campaigns by business intent and economics, not by vanity labels.

For most accounts, I’ll rebuild around a few “tiers” of intent. High-intent (ready-to-buy) traffic should have its own budget protection and tighter query controls. Mid-intent (comparison/solution-seeking) should be allowed to explore, but with stronger negatives and more educational landing pages. Brand (if applicable) should be separated because it behaves differently and can mask problems elsewhere.

If you’re using Performance Max alongside Search, you’ll also want to be deliberate about roles. Search can be your precision tool for known intent themes, while Performance Max can be used for broader incremental coverage—but only if conversion goals, assets, and business signals are clean enough to guide it.

Rebuild ad groups as “themes,” then write ads that clearly match the theme

Ad group design should make it easy to write specific ads and send traffic to the most relevant landing page. When ad groups contain mixed intent, you get generic ads, weaker expected clickthrough rate, weaker landing-page alignment, and you end up paying for relevance problems through higher costs and lower conversion rates.

A practical rule: if you can’t write a single responsive search ad that feels obviously perfect for every keyword in the ad group, your ad group is too broad.

While rebuilding, keep the number of “themes” manageable. Consolidation is often healthier than fragmentation because it gives the system more conversion data per decision point. The goal is clarity, not micro-management.

Use negative keywords strategically (and remember how negatives actually match)

A failing campaign is often a negative keyword failure. But the fix isn’t “add hundreds of negatives” at random—it’s building a repeatable negative strategy at the right level.

Negatives don’t behave like positive keywords. Negative keywords don’t automatically include close variants and expansions, which means if you want to exclude variations (plural/singular, synonyms, misspellings), you often need to add them explicitly. You should also use the right negative match type: negative broad (the default) blocks searches that contain all terms in any order, negative phrase blocks the exact sequence, and negative exact blocks only the exact query with no extra words.

When you restructure, I like to create a “shared logic” across the account: an account-level negative list for truly universal exclusions (jobs, free, definition, DIY—whatever is genuinely irrelevant), then campaign-level negatives to prevent overlap between intent tiers, then ad-group negatives for surgical control. Account-level negatives can be powerful, but use them cautiously: a single wrong exclusion can suppress good traffic everywhere.

Upgrade ads and assets before you judge the new structure

In modern Search, your ads are modular. If your responsive search ads are thin or repetitive, you’re asking the system to optimize with weak ingredients. At minimum, each ad group should have at least one responsive search ad with solid Ad Strength, and your headlines/descriptions should be genuinely differentiated (not 12 ways to say the same thing). Stronger ad relevance and stronger predicted engagement help you compete more efficiently because auction outcomes are influenced by more than just bid.

Also, don’t ignore creative enhancements that improve engagement. Adding image assets, a business logo, and a business name (where applicable) can lift performance because you’re giving users more information and more confidence at the moment of search.

Part 3: Reset bidding and budgets without blowing up learning (and without panic changes)

Respect Smart Bidding learning: restructure in stages, not all at once

If you’re using automated bidding, big structural changes can trigger learning. Learning duration is primarily driven by how many conversions the bidding system sees, your conversion cycle length (how long it takes users to convert), and the bid strategy type. As a practical benchmark, it can take up to around 50 conversion events (or roughly three conversion cycles) for a bid strategy to calibrate to a meaningful change. That doesn’t mean you can’t restructure—it means you should do it with intent, so you can attribute cause and effect.

When I restructure a failing campaign, I prefer a “controlled rebuild” approach: create the new campaigns alongside the old ones, migrate traffic gradually (by budgets, by keyword subsets, or by intent tiers), and only then sunset the legacy structure. This reduces the risk of a full-account performance cliff and gives you clean test windows.

Set budgets and targets that are mathematically compatible with your goal

A common reason campaigns fail is that targets are set like wishes instead of inputs. If you set a Target CPA that your market economics can’t support (given your conversion rate and expected CPC), the campaign will often under-serve or chase low-quality inventory. If you set a Target ROAS without reliable conversion values, you’ll get erratic decisioning.

If you’re starting fresh after a rebuild, consider beginning with a less restrictive target (or even a “maximize” strategy) to allow data accumulation, then tighten targets once you’ve regained stable volume. Budget also matters more than people think: if budget is too constrained relative to your goal, the system has fewer opportunities to learn and fewer auctions to choose from.

If you’re running a time-bound push (launch, event, short promo), consider using a time-bound campaign budget style where it fits your workflow, rather than manually yanking daily budgets up and down. The key is consistency: frequent budget whiplash can create noisy results and make it harder to diagnose what’s working.

Have a plan for conversion tracking outages and “weird weeks”

Campaigns often “fail” right after a site change, CRM outage, or tagging issue. If you’re on Smart Bidding and conversion data goes wrong for a period, the right fix isn’t guessing with bids—it’s protecting the bidding system from polluted data. That’s where data exclusions come in: they’re designed specifically for conversion tracking or conversion upload outages, not for excluding normal volatility or promotional spikes.

Separately, if you anticipate a short-term, unusual conversion-rate jump (like a flash sale), use a seasonality adjustment rather than trying to trick targets. These tools exist because bidding systems need context, and you’ll get cleaner recovery when you use the right lever for the right problem.

Use experiments to validate the restructure, not opinions

A restructure is a hypothesis: “If we group this intent together, align ads and landing pages, and set the right goal signals, performance will improve.” The cleanest way to validate that is controlled testing. Experiments let you measure impact while minimizing risk, especially when you’re changing bidding approaches, goal configurations, or major targeting logic.

When you test, keep it simple. Test one major concept at a time: new structure vs old, new bidding approach vs old, new messaging vs old. If you change structure, ads, landing pages, audiences, and bidding targets simultaneously, you might improve performance—but you won’t know why, and you won’t be able to scale the win reliably.

• Week 1: Fix measurement, clean goals, resolve policy/delivery issues, and map search terms to intent tiers.
• Weeks 2–3: Launch the new structure in parallel, migrate budget gradually, and stabilize ads/assets per theme.
• Weeks 3–6: Let learning settle (based on your tells: conversions and conversion cycle), then tighten targets and expand coverage using proven themes.

Part 4: What “good” looks like after the restructure (so you know you’re done)

Your account becomes readable, not just organized

The best restructure outcome isn’t a prettier campaign list. It’s an account where performance tells a clear story. You should be able to answer, quickly and confidently: which intent tier is profitable, which themes are scaling, which search terms are leaking waste, and which landing pages need work. When you can see that clearly, optimization stops being random tweaks and turns into deliberate improvement.

You can make changes without breaking everything

When structure, goals, and negatives are aligned, you’ll notice something important: small changes produce predictable movement. That’s the real sign you’ve successfully rebuilt a failing campaign. From there, scaling is straightforward—add budget to the best intent tiers, expand keyword coverage where search terms prove relevance, and keep feeding the system better assets and cleaner conversion signals.

When a Google Ads campaign is “failing,” the safest way to restructure it is to work in phases: first confirm measurement is telling the truth (clean up primary vs secondary conversions, values, sources, delays), then remove delivery blockers (disapprovals, overly tight budgets/targets/targeting), use real search intent evidence from the search terms report to identify waste and high-value themes, and only then rebuild around clear intent tiers and tightly themed ad groups with a deliberate negative keyword strategy and stronger ads/assets; after that, set budgets and CPA/ROAS targets that your economics can actually support, migrate changes gradually to respect Smart Bidding learning, and validate the new setup with experiments before fully switching over. If you want a lighter way to operationalize this without spending hours in audits and rewrites, Blobr connects to your Google Ads and runs specialized AI agents that translate best practices into concrete, reviewable actions—like refreshing RSA messaging with the Headlines Enhancer agent or improving message-match with the Campaign Landing Page Optimizer—so you can iterate on structure, queries, ads, and landing pages while staying in control of what changes and where.