One week your Google Ads campaigns generate sales. The next week, results drop—even though you have barely touched the account. If you manage an ecommerce store, that inconsistency can make it difficult to decide whether to increase budgets, pause products, or start over.
In a recent Google Ads Q&A, a viewer described helping a friend advertise an online shop selling mostly non-alcoholic beverages. The account had two campaigns: one for products priced above the market average and another for products priced below it. Some weeks were strong; others were disappointing. Here is how I would approach this problem, based on my experience managing PPC campaigns and the questions I raised in that discussion.
First: Are the fluctuations actually unusual?
Before changing the account, compare meaningful periods. If your store typically receives only a few sales a week, a difference of two or three purchases can make conversion rates and ROAS look dramatically different. Compare 28-day trends and similar days of the week, not just last Monday against this Monday. Allow for conversion lag: recent clicks may lead to purchases after your reporting date.
Google itself notes that campaign performance naturally varies and recommends checking campaign changes, bidding, budgets, conversion data, and auction conditions. See Google’s performance fluctuation guidance.
1. Review bidding strategy and conversion goals
Is each campaign using Maximize Conversions, Maximize Conversion Value, Target ROAS, Target CPA, or a manual bid approach? A restrictive target can limit participation in auctions; a strategy optimizing the wrong conversion can bring traffic that looks good in the dashboard but produces poor orders. Look at purchases, revenue, CPA, ROAS, and the actual business margin—not just clicks.
2. Check whether the campaigns are new or recently changed
New campaigns and bid strategies with little data can vary while building a performance history. However, not every weak week is a learning-phase problem. Google’s current Smart Bidding documentation explains that optimization is continuous and that budget or target changes do not automatically trigger the visible “Learning” status. Check the status and change history instead of guessing. Read about Smart Bidding learning.
3. Compare demand by date, location, and device
A beverage retailer may see changes in sales around weekends, paydays, weather, holidays, or local events. Treat these as hypotheses to test, not explanations you can assume. Segment your performance by week, day of week, location, and device. If conversion rates only decline on mobile, inspect the mobile checkout experience before changing bids.
4. Study competitor activity and your real offer
A competitor might launch a discount, reduce shipping charges, or enter the same auctions. Use Auction Insights where your campaign type supports it, then compare what shoppers receive from your store: product price, shipping, delivery times, stock, and trust signals. Auction Insights can reveal changes in competitive visibility; it cannot prove precisely which offer caused a sale to disappear.
5. Audit search terms and traffic quality
Where search-term reporting is available, compare relevant versus irrelevant queries in the strong and weak periods. Add narrowly targeted negative keywords when supported by the data. Avoid blocking valuable searches just to reduce the query count. For Shopping and Performance Max, also review the product-level and search-insights reporting available for that campaign type—visibility is not identical across all campaign types.
6. Question the above-market versus below-market campaign structure
Separating premium-priced products from competitively priced products can be sensible if the products, margins, bidding goals, and targeting are genuinely different. But price alone is not always the best division. One campaign could have higher-value orders but lower conversion rates; the other could generate more orders with thinner margins.
Look at the exact product IDs included in each campaign. Avoid unintended overlap or internal competition, and do not assume the seller is offering the same item at two prices merely because the campaigns are labeled differently. If using a Merchant Center feed, custom labels can help organize products by price position, margin, or seasonality. Google’s guide to Shopping custom labels explains the options.
7. Verify conversion tracking before optimizing
Compare Google Ads purchase data with your ecommerce platform and GA4, remembering that the systems may attribute conversions differently. Check for duplicate purchase events, missing transaction values, broken checkout tags, changes to the primary conversion action, and consent-related reporting changes. A measurement fault can look like a sudden sales collapse.
8. Inspect product data and landing pages
For ecommerce campaigns, check Merchant Center availability, feed disapprovals, price mismatches, stock status, shipping settings, and product titles. Review the landing page on real mobile devices. Slow loading, unclear delivery fees, broken payment options, or out-of-stock bestsellers can damage results without any Google Ads setting changing. Google’s Shopping monitoring guide details useful product-level checks.
9. Use a controlled testing plan
Rather than changing keywords, targets, budgets, and pages together, identify the biggest performance change and test one logical explanation at a time. If impressions dropped, investigate eligibility and auctions. If clicks held steady but purchases fell, investigate traffic intent, checkout, pricing, and measurement. If order value changed, inspect the product mix and promotion.
What I would ask the store owner before making a change
- Which campaign type and bidding strategy is being used?
- How many conversions and how much revenue does each campaign generate over 28 days?
- What changed in the account, store, stock, shipping, or promotions?
- Are the two campaigns advertising distinct product groups, and what margins do they generate?
- What do search terms, Auction Insights, product performance, and conversion tracking show?
My final advice
Inconsistent Google Ads results are a symptom, not a diagnosis. I would first review conversion data, bidding settings, demand patterns, competitors, search terms, product groups, and the landing page before rebuilding a campaign. A stable testing process usually teaches you more than reacting to every poor week.
Need a second opinion? My team at AARSWEBS offers an initial free Google Ads audit. We can review your setup and identify practical priorities. Any implementation or ongoing management is quoted separately, and results cannot be guaranteed.
About the author: Ali Raza, digital marketer and founder of AARSWEBS Solutions Co LLC, a Google Partner agency.
