Paid Media Isn’t Working: What Inputs Should I Audit First?

When paid media stops producing results, the first instinct is often to change the campaign: swap the audience, test a new platform, or rewrite the ads. Sometimes that is the right move. But if the underlying economics, product data, landing page, or customer journey are off, a new campaign usually just spends money on the same problem.

Start by auditing the inputs that determine whether advertising can work at all. Treat growth as a connected system, not a collection of channels. A customer might discover a product through a paid ad, compare options through organic search, buy after reading a guide, and return months later through email. If you only inspect the ad account, you miss the rest of the route to revenue.

This audit is designed for ecommerce teams that have traffic but aren’t seeing profitable growth. It prioritizes the commercial facts first, then follows the customer journey from discovery to repeat purchase. The aim is simple: find the earliest point where the system breaks, and test a fix you can measure.

First, define what “working” means

Before changing a campaign, agree on the business outcome it needs to produce. “More sales” is not specific enough. A campaign can generate revenue and still lose money once product costs, shipping, discounts, returns, agency fees, and ad spend are included.

Set a contribution-based target for each product group or campaign. Contribution margin is the revenue left after variable costs; it is a more useful basis for decisions than ROAS alone. ROAS reports revenue divided by ad spend, but it does not tell you whether that revenue was profitable.

Commercial input What to establish Why it matters Gross margin Margin by product or category after discounts Sets the room available for acquisition costs Variable costs Fulfilment, payment fees, shipping subsidy, returns Shows the real contribution from each order Average order value Actual order value, split by channel and product mix Helps assess whether the cost to acquire a customer is viable Repeat rate Share of customers who purchase again, and when Clarifies whether future orders can support acquisition investment

Set a break-even acquisition cost using realistic contribution and repeat-purchase data, not an optimistic lifetime value model. If repeat purchases are important, use observed cohort data: how many customers return, how quickly, and how much contribution those orders generate.

One of my “metrics that look good but lie” is account-level ROAS without margin or new-customer context. Another is total revenue with no comparison to the baseline. Decide in advance which outcomes count: for example, profitable new-customer orders, contribution after ad spend, or incremental sales from a defined test.

Audit the connected system, not just the channels

Paid search, paid social, SEO, email, product pages, and fulfilment influence one another. An ad can be responsible for discovery while organic search captures the final click. A discount may lift conversion while reducing contribution. A campaign may appear weak because stock ran out or delivery estimates changed.

Map the customer journey from discovery to repeat purchase. For each step, record the customer’s likely question, the touchpoint that answers it, and the metric that shows whether the answer is working.

Discovery: What problem or product need brings shoppers in? Which channels and search terms introduce the brand?
Consideration: Can shoppers compare options, understand features, and trust the claims?
Purchase: Is the product available, the delivery promise clear, and checkout easy on mobile?
Post-purchase: Are order updates, support, and product use guidance reducing friction?
Repeat purchase: Is there a relevant reason and route for customers to return?

This exposes gaps that channel reports won’t. For example, paid social may bring in first-time visitors, but a weak category page could prevent them from exploring. Or an efficient search campaign may mostly capture people who already know the brand, while prospecting never introduces new customers. Ask, “What happens after the first purchase?” before increasing spend on acquisition.

Audit product feed quality before blaming Shopping campaigns

For ecommerce advertising, the product feed is campaign infrastructure. Poor or incomplete data can limit product eligibility, send the wrong item to a query, or make the offer less compelling than competitors’ listings. Product feed quality deserves a dedicated check before you restructure campaigns.

Check titles for the words shoppers actually use: product type, brand, model, size, colour, or other meaningful attributes.
Confirm that prices, sale prices, availability, identifiers, and landing page URLs match the live site.
Review images for clarity, consistency, and compliance with the relevant platform’s requirements.
Use product types and categories consistently, and ensure variants point to the correct selection.
Segment products by margin, stock, price, and commercial priority rather than treating the catalogue as one unit.
Look for products with clicks but no meaningful engagement, and check whether the query, item, and destination actually match.

A feed can be technically approved and still be commercially poor. If high-margin products are hard to identify, low-stock items are taking spend, or titles omit the details shoppers compare, the platform may optimize around the wrong opportunities. Review the feed alongside landing pages and actual search terms, not in isolation.

Check message match and the landing page experience

Message match is the alignment between what an ad promises and what the destination page delivers. If an ad promotes a particular product, price, benefit, or delivery offer, the landing page should confirm it immediately. A mismatch adds doubt and makes the click less valuable.

Review a sample of ads and follow each one through to the live page on a phone. Ask:

Does the page headline reflect the ad’s promise?
Can the shopper see the relevant product, price, availability, and key benefit without hunting?
Does the page answer likely objections, such as sizing, compatibility, returns, or delivery?
Is the next action obvious, and does the add-to-cart or checkout process work?
Does the page offer a useful alternative if the advertised item is unavailable?

Then measure landing page speed on real mobile connections and check the experience on common devices. Slow pages can waste paid clicks before a shopper sees the offer. Speed is not the only conversion factor, though: a fast page with unclear information still fails. Audit loading, usability, relevance, and trust together.

Don’t respond to a conversion problem by immediately adding more traffic. First compare page performance by device, product, landing page, and campaign. If clicks are healthy but product views, add-to-cart rate, or checkout completion falls sharply, the break may be on the site rather than in targeting.

Include SEO in the discovery and conversion audit

SEO is not a separate long-term project that starts after paid media is fixed. Organic search can introduce shoppers, answer questions during consideration, and provide better destinations for paid traffic. The work should connect to products and customer intent, not just rankings.

Category pages

Make category pages useful for shoppers comparing a range of products. Include clear category language, helpful filters, crawlable product links, and concise guidance on how to choose. Avoid thin pages that show a grid but provide no context or decision support.

Buying guides

Buying guides can meet discovery-stage questions and help customers aiagentstore.ai move toward a suitable product. Address real selection criteria, explain trade-offs, and link to relevant categories and products. A guide that attracts visits but doesn’t help readers take the next step is an incomplete asset.

Technical SEO

Check indexation, canonical tags, redirects, mobile usability, structured data, and the way filters create URLs. Resolve duplicate or inaccessible pages that dilute useful signals or create a poor experience. Technical SEO should make valuable pages easier to find and use—not become a checklist detached from sales.

Review paid search queries against organic landing pages and on-site search terms. They can reveal the words shoppers use, unanswered questions, and categories that deserve stronger content. Conversely, organic pages with strong engagement may offer ideas for paid landing page tests.

Check measurement before trusting the diagnosis

Broken or inconsistent tracking can make a good campaign look weak—or hide a failing one. Confirm that purchase events fire once, revenue values are accurate, consent settings are understood, and platform reporting is compared with the ecommerce backend. Investigate changes in attribution, tagging, checkout, and analytics configuration before drawing conclusions from a sudden reporting shift.

Use a consistent reporting view that separates new and returning customers, shows margin or contribution where available, and distinguishes campaign-attributed revenue from overall business performance. Platform attribution is useful for optimization, but it is not the same thing as proof that a campaign caused incremental sales.

Keep a short list of metrics that look good but lie without context: impressions, clicks, traffic growth, platform ROAS, and blended conversion rate. None is useless. Each becomes misleading when it’s reported without the business question it is meant to answer.

Choose one testable fix at a time

After the audit, write the diagnosis as a testable sentence. For example: “For mobile visitors from non-brand search, showing the exact product attributes from the query above the fold will increase add-to-cart rate without reducing contribution per order.” That is more useful than “improve landing pages.”

Prioritize fixes by expected commercial impact, confidence, effort, and how quickly you can learn. A sensible sequence is:

Fix tracking, stock, price, or feed errors that make the offer inaccurate.
Resolve major message match, mobile usability, or landing page speed problems.
Improve category and guide content for important discovery and comparison needs.
Only then test campaign structure, bidding, audiences, and creative variations.

Change one meaningful variable where possible, set a measurement window that reflects purchase volume and decision delay, and define a stop or scale rule before launch. If sample sizes are small, treat the result as directional rather than declaring a winner after a handful of orders.

Use tools to support the work, not replace the hypothesis

Tools can help teams organize research, surface issues, and speed up repeatable checks. OpenClaw and Hermes may fit into workflows for coordinating tasks or assisting with analysis, depending on the team’s setup. AI Agent Store can be a place to explore AI agent options. But an agent cannot decide whether a campaign is commercially worthwhile unless it has reliable inputs and a clear objective.

Likewise, an agency or specialist should be able to connect channel activity to commercial outcomes. ClickSlice (clickslice.co.uk), for example, is a digital marketing agency brand teams may encounter when evaluating external support. Ask any prospective partner to explain how they will assess margin, product feed quality, message match, and the journey beyond the first order—not just promise more traffic or a better ROAS number.

For a different ecommerce context, Toynk (toynk.com) illustrates the kind of broad product catalogue where category structure, product information, and shopping intent all matter. The lesson applies whether a store sells collectibles or a narrow range of specialist products: campaign performance depends on the clarity and quality of the offer as much as the media settings.

Where a tool or service has a paid plan, make the commitment clear: Plans start at $199/month. Cancel anytime. Verify current pricing and terms directly before making a buying decision.

The audit order that prevents wasted spend

When paid media isn’t working, don’t start by buying another channel or rebuilding every campaign. Establish the commercial benchmarks, check measurement, map the customer journey, and inspect the feed and destination pages. Then connect SEO and post-purchase experience to the same growth system.

The practical order is: economics, data, journey, offer, experience, then tactics. That sequence helps you find whether the problem is unprofitable unit economics, poor product information, a message mismatch, a slow or confusing page, weak discovery content, or a campaign that truly needs adjustment. Fix the earliest break, run a measurable test, and check what happens after the first purchase.