Measuring Paid Ads the Right Way in 2025

By far the most common frustration I see with performance marketing teams is confusion with the question “is it working?”. You spent $100K on paid ads or maybe even just $1K, but how do you know for sure how much you got back?

Short answer, there’s no one way that will give you the whole truth. Instead you have to look in a few places. Today we’re going to cover what those places are and how anyone can and should know if their paid spend is working with high confidence.

For some context setting, let’s first cover all of the different ad channels out there. Each one has it’s own purpose and position in your Marketing funnel. Starting with Awareness, then Consideration, then Conversion. Each stage is valuable it’s own way, but each with a different job than the next.

🔼 Top of Funnel (Awareness)

These are great for broad reach, cold audiences, and brand visibility.

• Meta Ads (Facebook, Instagram) – excellent for interest-based targeting and content discovery

• TikTok Ads – viral-style creatives, best for native content and cold traffic

• YouTube Ads (Skippable/In-stream) – high reach, strong visual storytelling

• Display Ads (Google Display Network, programmatic) – broad banner placements

• Snapchat Ads – youth-focused, short-form mobile video

• Pinterest Ads – visual discovery platform, good for lifestyle/ecomm brands

• Connected TV (Hulu, Roku, etc.) – non-clickable, used for large awareness pushes

• Reddit Ads – niche communities, contextual targeting

• Podcast Ads (programmatic or host-read) – strong for affinity and trust

• Spotify Ads – audio and video formats for broad reach

• Digital Out-of-Home (DOOH) – billboard-style screens in public places

• Conferences – hosting or sponsoring, can be lower funnel if nurturing existing leads

➡️ Middle of Funnel (Consideration)

These help nurture interest, drive retargeting, and show product value.

• Retargeting (FB/IG) – recapture visitors who didn’t convert

• YouTube In-Feed Ads (Discovery) – show up in search results & recommended videos

• Google Discovery Ads – native placements across Gmail, YouTube, and GDN

• LinkedIn Ads – good for B2B consideration, especially lead gen

• Twitter/X Ads – effective for engagement and topic-based targeting

• Quora Ads – context-based, high-intent Q&A targeting

• Programmatic Display (retargeting or behavioral) – reinforces message across the web

• Native Ads (Outbrain, Taboola) – article-style placements on publisher sites

• Pinterest Consideration Campaigns – nurture users closer to purchase

• Email Sponsorships (e.g. The Hustle, Morning Brew) – direct reach to curated audiences

🔽 Bottom of Funnel (Conversion)

These are best for high-intent users, retargeting, and direct-response goals.

• Google Search Ads (PPC) – extremely intent-based ("buy now" searches)

• Bing/Microsoft Ads – same as above, less competition

• Google Shopping Ads – product-based, direct to purchase

• Meta Advantage+ Shopping Campaigns – optimized for conversions and purchases

• Retargeting on all platforms – especially Dynamic Product Ads (DPA) on Meta or Performance Max (Pmax) on Google

• Affiliate Marketing (paid placement) – pay per conversion or lead

• SMS/Email remarketing – via paid list rental or cold outreach tools

• Amazon Sponsored Products – for marketplace sellers, very high buying intent

• Instacart/Drizly/UberEats Ads – high intent for food/bev brands

• Retail Media Networks (Walmart, Target, CVS, etc.) – search and display ads on retailer sites

So many channels, so little time. A lot of Marketers won’t use all of these, with most of the ad dollars going to Meta Ads (Top/Middle Funnel) and Google Ads (Middle/Bottom Funnel). Going back to the topic of this post, how does measuring these channels change depending on the stage they’re in? Well, a lot. The further up the funnel you go, the harder it becomes to measure the value. Why? Those people are not usually going to convert on the same day they see the ad. You’re building awareness or consideration, then relying on other channels further down to convert them into a customer.

That doesn’t mean one channel is better than any of the others. Like I mentioned each has it’s own job. The Seven Touches of Marketing principle tells us that on average someone needs to see your brand 7 times before buying.

Of the 7 touches, Meta ads (FB and IG) are usually one of the first touch points. Why? People are minding their own business doom scrolling and your ad pops up. They weren’t searching for it so chances are they’re not going to buy your product right away. It introduces people to your business and gets them thinking about purchasing your product, then when they visit your website you follow them around the internet with the middle and bottom of funnel channels till they buy.

Let’s go deeper into Meta ads. Everybody’s doing it. It’s still the hot channel. So as a self-respecting paid Marketer how could you not try it? If so many people are doing it, surely they can prove the value they’re getting out of it right?

I’m always shocked with how many ad dollars go to Meta yet no one seems to have a good answer to prove if it’s actually helping their business. Let’s take a real life example of an ecomm store that spends 6 figures/month on Meta ads. This is an actual month’s worth of data from from December 2024 using 4 popular analytics tools. Google Analytics, Shopify, Northbeam, and then within the ads platform in Meta Ads Manager.

The actual spend $189,648.03. That’s how much left the bank account but how much did we actually make from that? Let’s see.

Google Analytics 4 says we made $217,491.91 in revenue.

Shopify Analytics says we made $98,243.89 in revenue.

Northbeam says we made $56,617.71 in revenue.

Meta Ads Manager says we made $502,977.95 in revenue.

WTF why are these numbers so different and which one is right? Well the truth is they are all right at what they’re measuring. Not helpful. This is what makes top of funnel channels so difficult to do as a Marketer. Everyone who’s managed paid ads knows the pressure of your boss or CEO constantly grilling you to prove what you’re doing is growing the business so most people (especially agencies) just take the highest number of the 4. Of course that will always be Meta itself reporting that their ads are doing the best because they want you to spend more money.

But how can they say such a different number than the others? The answer lies in how each platform tracks or attributes the sale. Here’s a general outline of each of the attribution models.

Meta actually goes one step further. The default attribution they use is tracking “Any click” that leads to a sale in the next 7 days (they follow you around with a cookie) AND anytime someone even sees your ad then converts in the next 24 hours. That’s taking credit for a lot and also you’re paying for every one of those clicks and views. Meta charges you on a CPM or cost per thousand impressions. The “view” part always gets me because of how aggressive Meta gets to count a view impression with any part of the cookie appearing on a users screen.

In my opinion I think this is taking too much credit to say they were the reason that person made a sale and 100% is attributed back to the ad they saw on FB. In reality that person probably interacted with your brand several more times across different channels that deserve some of the credit as well.

So if that FB ad is how your eventual customer first heard about you, but then they actually bought your product from your email or affiliate, that FB ad has to be worth something right? Definitely. It should take a % of the credit but how much depends on how many other touch points happened, if there were any.

So here’s my recommendation. First look at the best last click analytics channel you have. For this example it would be Shopify analytics, it’s the platform where your customers are checking out with your product so they will have solid last click recommendation. In our case Shopify is saying we made $100K on a $190K spend or a 0.53:1 return on investment at the very minimum of value. It’s at least driving that much in directly attributable revenue. But there’s also the assisted revenue (before the purchase) and repeat revenue (LTV) to consider. Some tools that can help you with this that I’ve personally used are Hyros, Wicked Reports, or Rockerbox. Don’t overcomplicate it, pick one that can give you reliable last click then move on.

But as we discussed that doesn’t tell the full story since most of Meta’s conversions are first click not last click. To complete the rest of the puzzle we need to take data into our hands with a post-purchase survey, which is an excellent way to get directional data to attribute sales back to the original marketing source. It won’t be 100% accurate but will help you understand the true effect of top of funnel channel like Meta. They can be very simple and are super easy to set up. A few tools I like are KnoCommerce, Fairing, or if you want to do more custom SurveyMonkey and Typeform.

At AppSumo we track post purchase transactions by channel in a Google Sheet alongside the last click transactions that we record ourselves through UTMs stored in our own database (Snowflake). As you can see below, the number of transactions we’re capturing as last click vs the survey can be wildly different. This is exactly why you need to do both. As you continue to diversify your marketing mix the path to conversion will become messier, but the role of proving top of funnel channels like Meta ads becomes that much more crucial.

Going the extra mile to look at the full picture can be annoying and nuanced at times, but necessary if you want to prove the true value of your paid marketing. No one tool will be the single source of truth, instead take each with a grain of salt and capture as much data on your own as you can to make an informed to decision on how to spend.