A campaign bidding the same amount for every impression regardless of where the user is or what device they are on treats all traffic as equal. It is not. A click from a desktop user in Germany and a click from a mobile user in Brazil do not cost the same, do not convert at the same rate, and do not produce the same revenue. Treating them identically means overpaying for one and underpaying for the other, and the gap between the two grows with every dollar spent.
Most ad platforms allow advertisers to adjust bids by device type and geographic region. These adjustments are percentage modifiers applied on top of a base bid. If mobile traffic converts at 60% of the desktop rate, a negative bid adjustment on mobile reduces the price paid per impression to reflect that lower return. If a specific country delivers conversions at half the cost of another, a positive adjustment on that country captures more of that inventory. The adjustments are simple in isolation. Where they get interesting is in what happens when they stack up over weeks and months.
The Device Split
An analysis by DRIP Agency covering 486 million sessions across 117 European e-commerce brands (March 2025 to February 2026) found that desktop converted at 1.56 times the rate of mobile, while mobile accounted for 78% of total traffic. That pattern is common across most verticals: mobile sends more visitors, desktop closes more sales.
In Byron Sharp’s How Brands Grow, the core argument is that brands grow primarily by increasing their reach across the full buyer base rather than by targeting heavy buyers more intensely. Applied to device bidding, this means a campaign that bids heavily on desktop because desktop converts better may be efficient in the short term but is narrowing its reach to a shrinking share of the audience. Mobile is where the majority of impressions live. A bid adjustment that prices mobile too low cuts the campaign off from the largest pool of potential buyers.
The compounding effect works in both directions. A 20% negative adjustment on mobile, applied daily over 30 days, reduces mobile delivery by roughly the same proportion each day. If mobile conversion rates improve during that period because the landing page was updated or a checkout issue was fixed, the campaign misses that improvement because the bid adjustment was set based on last month’s data. Bid adjustments that are not re-evaluated regularly lock in old performance assumptions.
The Geography Split
Geographic bid adjustments have larger variance than device adjustments. CPM costs between countries can differ by a factor of ten or more. TrafficJunky’s network, which serves traffic globally across multiple sites, shows the same pattern: impressions in Tier 1 countries like the United States, Canada, and the United Kingdom cost significantly more than impressions in Latin America or Southeast Asia.
The question is whether that cost difference is proportional to the value difference. A campaign paying $8 CPM in the US and $0.80 in Brazil is paying 10 times more per thousand impressions. If the US converts at ten times the rate of Brazil, the cost per conversion is the same and the geographic split is neutral. If the US converts at only four times the rate, the advertiser is overpaying for US traffic relative to what it produces. Without separate tracking by geography, there is no way to know.
How Small Adjustments Compound
A $5,000 daily budget with a 15% positive bid adjustment on desktop and a 20% negative adjustment on mobile shifts roughly $600 per day from mobile to desktop. Over 30 days, that is $18,000 redirected based on a single assumption about device performance. If the assumption was correct at the time it was set, but the underlying data has shifted, the campaign has spent $18,000 reinforcing an outdated pattern.
Geographic adjustments compound the same way. A 25% positive adjustment in the United States and a 30% negative adjustment in a lower-cost region change the geographic mix of the campaign significantly within a week. Layering device and geography adjustments together multiplies the effect. A US desktop impression might receive a combined uplift of 40% or more, while a mobile impression in a Tier 3 country might be bid down by 50%. The combined adjustment determines where the money actually goes, and the further those adjustments drift from current performance, the more waste accumulates.
Keeping Adjustments Current
The fix is not to avoid bid adjustments but to treat them as variables that expire. A reasonable cadence is to re-evaluate device and geographic modifiers every two weeks for campaigns spending more than $1,000 per day and monthly for smaller campaigns. Each review should compare the adjustment in place against the actual cost per conversion for that segment over the most recent period. If the data no longer supports the modifier, the modifier needs to change. A bid adjustment is a hypothesis about where the budget will perform best. Like any hypothesis, it needs to be tested against new data.