The hidden cost of bad traffic data

Most teams already account for the wasted media spend. The bigger problem is that bad traffic has already skewed every decision the team makes off this reporting. Budget shifts toward an audience that looked strong and turned out to be bots. Next quarter’s plan gets built on a read of this quarter that was never true, and the gap between what the reports say and what’s actually happening keeps growing.

‍

Campaign inefficiency that gets harder to explain. CPAs creep up. Audiences that used to deliver somehow don’t anymore. The team is doing the work right, but it’s basing all of it off bad traffic data.

‍

Why it stays invisible

‍

These problems go uncaught because most teams don’t go back and analyze campaigns that appeared successful. Once a campaign hit its acquisition goal, the auditing stops. Then a few months later, when a competitor with cleaner data is reaching the same buyers, or the CPA drifts up because too much of the spend is going to bots, flags start getting raised.

‍

Flags lead to questions, but often it’s the wrong ones. They’re too easy. The problem only surfaces if you go back to the conversion data and ask hard questions. Like “Of the people who actually did what the campaign was built to drive, how many came from the audience we targeted?  And of the people in the targeted audience, how many ever came anywhere close to converting?”

‍

Bad traffic isn’t just coming from bots

‍

Bots and click farms account for a meaningful percentage of digital traffic. That portion was never human to begin with.

‍

But bad traffic also includes real humans reached at the wrong moment. A record that was accurate two years ago has long since expired. Someone who was in the market for a vehicle eighteen months ago, bought one, and has no reason to see another auto ad for six more years. A click from that person counts the same in the dashboard as a click from someone still shopping.

‍

Stale targeting and fraudulent clicks cause the same problem. Both put people at the top of the funnel who have no path to the bottom. Both produce reports that look clean because the dashboard can’t tell good traffic from bad.

‍

The scoring algorithm problem

‍

Most teams rely on off-the-shelf tools to filter bad traffic out of their reports. Those tools score it, but the algorithms behind the scoring aren't disclosed. You don't know what they consider valid, how often they update their criteria, or whether their definition of a bot has kept pace with how sophisticated that traffic has gotten. You get a number and take it on faith.

‍

Building something better requires having enough data to know what real traffic actually looks like. Site Impact's database covers more than 200 million U.S. consumers across 750 attributes, the depth needed to understand what separates a genuine human visitor from traffic that isn't. Because the off-the-shelf tools couldn't give us that standard, we built our own. The bar we hold is that a real human has to make it to the offer. Fraudulent and bought traffic is an industry problem, not a Site Impact problem, but if we're going to deliver real value to clients, we need to make sure what we're sending actually reaches what the campaign was built for.

‍

What’s worth measuring instead

‍

A lot of campaigns optimize around top-of-funnel metrics like clicks and impressions. The metrics that tell you whether a campaign actually worked live further down. Scheduled appointments. Completed forms. Test drives booked. Home tours arranged. Signups to see the demo. These require a human being to decide to take an action, which is harder to fake and harder to inflate.

‍

When the top-of-funnel numbers look good but the bottom-of-funnel numbers don’t match, the traffic itself is the first thing to audit. Not the creative. Not the offer. Not the landing page. If the data coming in is unreliable, nothing downstream of it will fix it.

‍

Before the campaign goes live

‍

The time to audit the traffic is before the campaign launches. Start with the tool doing the scoring, the service that rates each click as real or not before the campaign counts it. Know what yours is letting through, and why. Know what your downstream conversions actually require—Is it clicks anyone (or anything) can produce, or actions a person has to commit to? Know whether the audience is the same one you built eighteen months ago.

‍

The campaigns that hold up when you go back to check them are the ones where that work happened before the go-live. The ones that don't usually skipped it.