How to measure a change in your own campaigns
Save the starting figures, record what you change and compare similar periods afterwards. Include genuine enquiries or paid orders alongside traffic counts so you can judge whether the result helped your business.
Read moreRead less
Note promotions, new ads and website changes during the comparison. Several adjustments can affect the same result. The examples here help you plan an investigation; published customer results would need their own dates, methods and supporting records.
What makes a before-and-after comparison useful?
Look for a clear starting point when reading a case study: the business, advertising channels, dates and problem being addressed. Check what changed and which result was measured. Fewer suspicious visits, lower advertising charges and more qualified sales describe different outcomes. The report should make clear which one actually improved.
Ask whether the business also changed its ads, budget, offer or landing page during the same period. Those changes may affect the result. If you want to run your own comparison, save the original campaign report and note each adjustment as it happens. Choose similar days and allow for the time customers normally take to buy. Use a relevant case study to prepare questions for your trial, especially about setup and reporting. The scenarios and diagrams here explain how to investigate a problem; they are not published customer performance results or a prediction of the saving your account will achieve.
What should you measure before switching protection on?
Before changing protection settings, save a report for the campaign you want to improve. Include its spend, clicks, landing pages and useful results: qualified calls, accepted enquiries or paid orders. Note the existing exclusions as well. Choose dates that reflect your normal business. A weekend promotion will usually bring different visitors from an ordinary Tuesday.
Keep a short note of changes during the trial, including new ads, budget increases, promotions and website edits. If sales improve after several changes at once, that note will help you understand what may have contributed. Check lead quality with the people answering calls or handling enquiries; a larger form-submission count can conceal more unsuitable requests. With a low-volume campaign, you may need longer to judge the effect on sales. You can still use the first week to confirm that tracking works, reports make sense and the right account is connected.
Why do ad clicks and website visits show different totals?
Your advertising account and website analytics count different things. A visitor can click twice during one session, leave before a slow page loads or decline analytics cookies. A redirect can also remove campaign parameters. Start by checking these possibilities when the numbers do not match; the difference alone cannot tell you how many clicks were fraudulent.
Use the same date range and time zone in both reports. Check that you selected the same campaign and landing page, and allow for reporting delays. Open the actual destination URL and follow any redirects to see whether its campaign information is kept. If an enquiry becomes a sale several days later, include that delay when reviewing results. For recurring discrepancies, save a specific example with the click time and destination. It will give the person maintaining your tracking a much clearer starting point than a screenshot of two unexplained totals.
How much money has click protection actually saved?
Start with what you were actually charged and the results your business received. A flagged visit does not automatically equal a paid click. Your advertising platform may already have filtered or credited some invalid activity. Check its reporting and billing adjustments before adding up a claimed saving from the same visits.
For a practical comparison, look at spend per qualified enquiry or paid order before and after the change. Include the subscription fee and the time you spend reviewing traffic. For example, if ten flagged visits are multiplied by a $4 average click price, the resulting $40 is an estimate of possible exposure, not proof of a $40 refund. Keep estimates labelled and explain the calculation. A useful review should show whether your account is producing better-quality enquiries or orders for its overall cost, with promotions, bid changes and other adjustments taken into account.
How to investigate a sudden increase in ad clicks
Suppose a campaign suddenly receives more clicks without more enquiries. Before investigating, you only know that costs have risen. Pausing the campaign may stop the spending, but it can also stop genuine customers reaching you. Start by noting the affected dates and checking whether you changed the budget, targeting, offer or landing page around that time.
Next, inspect the traffic source and a few visits from the affected period. Test the enquiry form yourself. If you find a page problem, repair it. If a particular group of visits looks suspicious, review the available protection settings for that campaign and record the change you make. Afterwards, watch both the suspicious activity and genuine enquiries. This before-and-after example shows how to narrow the response. To assess your own result, compare similar days and keep other campaign changes in the record so you can see what helped.
Are your enquiries reaching the right customers?
Ask the people answering calls or following up forms what they are receiving. Are enquiries genuine? Is the customer in your service area? Do they want a service you provide? A real person asking for something you do not offer is a targeting problem. Repeated nonsense submissions may need a closer look at the form and associated visits.
Keep a simple record of the reasons enquiries are rejected. It can reveal a broad keyword, an unclear offer or a form that is too easy to misuse. Check duplicates before dismissing them: a customer may have submitted twice because no confirmation appeared or nobody responded. Where the visit looks suspicious, compare its timing and campaign with the submitted details. Use both the traffic information and feedback from your team when deciding what to change. Improving an ad, clarifying a service area or fixing follow-up can be as relevant as adjusting a protection rule.
Why are shoppers clicking but not placing orders?
Customers often visit a shop more than once before buying. They compare sizes and prices, check delivery costs or return after payday. Start with the products and traffic sources behind the change in sales. Are you advertising an item that is out of stock? Does the landing page show the same price and delivery promise as the ad? Are mobile visitors getting further than the product page?
Try adding an item to the basket and completing the checkout steps yourself. Check shipping restrictions, payment errors and any recent changes to tracking. Once those basics work, inspect the visits that still look unusual, such as repeated activity concentrated on one product or source. Compare them with paid orders from the shop's own records. An abandoned basket, a suspicious ad visit and a disputed payment each need a different investigation. Identifying which problem you have helps you choose a useful response without interrupting genuine shoppers.
What changes during a sale or product launch?
A sale, press mention or product launch can bring a large increase in genuine visits. New visitors may browse differently from regular customers, and campaigns may reach audiences that rarely appeared before. Note when the promotion started, which products it featured and any budget or targeting changes made alongside it.
When inspecting unusual activity, compare the same source, device and product group where possible. Check that the advertised items remain available and that checkout can handle the extra demand. If customers report problems after a protection change, investigate promptly with the person who manages the site. Once the promotion ends, compare paid orders or qualified enquiries with the spend and traffic from that period. This gives you a useful record for the next sale. It also helps you distinguish a genuinely successful campaign from a tracking error or a burst of visits that did not reach the intended customers.
Does a click fraud statistic apply to your business?
Before using a statistic in a budget discussion, check its source, dates and definition. Was the study counting ad impressions, clicks, website visits or requests? Which channels and countries did it cover? A report about one type of traffic may tell you little about a local service campaign or a specialist B2B business.
Keep the link and the definition with any number you quote. If a report gives a percentage, check what that percentage is a share of and how the researchers identified invalid activity. Then compare the question with your own account. You might inspect a particular network or source because a report highlights it, but use your campaign records to judge its effect on your business. For a client presentation, explain these limits in ordinary language. A useful statistic helps you decide where to look; it cannot tell you what a particular advertiser has already lost without account-specific information.
What should a useful traffic report show?
Start your report with the campaign, dates and question you investigated. For example: enquiries fell after a new placement started sending traffic. Show the change in spend and qualified enquiries, then add a few visits that explain why you investigated that source. Include the action taken and any question that still needs an answer.
If the report goes to a client, explain unfamiliar terms the first time you use them. They should be able to see what happened without interpreting a list of IP addresses. Keep the detailed records available for the person checking the account. Where a change has been made, say when you will review its effect and what you will compare. A short report with a specific follow-up is easier to use in a campaign meeting than a long list of blocked visits with no explanation of how customers or advertising costs were affected.
