RPM vs CPM explained simply

CPM is not your paycheck

CPM stands for "cost per mille" (cost per thousand) and usually refers to what an advertiser pays for one thousand ad impressions. It is the buy-side metric—the price the brand or agency pays to have their ad shown. CPMs can range from $1 to over $50 depending on the industry, targeting, and ad format.

When creators quote CPM as if it were their earnings, they often dramatically overstate what they can expect to keep. The advertiser might pay a $20 CPM, but after the platform takes its share (YouTube keeps 45%, for instance) and after accounting for the fact that not every view generates an ad impression, the creator's actual take-home is significantly lower.

Using CPM as your planning number is like using a company's revenue when you want to know an employee's salary. It is a related number, but it is the wrong one for your purposes.

RPM is usually the better planning number

RPM stands for "revenue per mille" (revenue per thousand views) and represents what you actually earn per 1,000 views after the platform has taken its share. This is the number that directly translates to your income: multiply your RPM by your view count (in thousands) and you have your estimated earnings.

YouTube reports RPM directly in your analytics dashboard. For other platforms, you may need to calculate it yourself by dividing your total earnings by your total views and multiplying by 1,000. That is why the ContentPaycheck calculators let you edit RPM directly and even switch platforms, instead of forcing one preset number.

RPM varies widely. A personal finance YouTube channel might see an RPM of $12-25, while a gaming channel might see $2-5. A TikTok creator might see effective RPMs below $1 from the Creator Fund. Understanding your actual RPM is essential for honest income planning.

Why RPM changes over time

Your RPM is not a fixed number. It fluctuates based on several factors: the time of year (Q4 is highest due to holiday ad spending), your audience demographics (US/UK/Canadian viewers generate higher RPMs than viewers from countries with lower ad markets), your content niche, and the overall state of the digital advertising market.

Watch time also affects RPM indirectly. Longer videos can include more mid-roll ad placements, increasing the number of ad impressions per view. A 20-minute video with three mid-roll slots will typically generate a higher RPM than a 3-minute video with a single pre-roll ad.

Seasonal patterns are predictable enough to plan around. January RPMs are often 30-50% lower than December. Smart creators use this knowledge to time their most important content releases and product launches.

Platform differences matter

RPM varies dramatically across platforms because of differences in ad formats, audience intent, ad inventory supply and demand, and platform revenue-sharing models. YouTube generally offers the highest RPMs for long-form video because it has the most mature ad ecosystem and supports multiple ad formats per video.

TikTok's Creator Fund has historically offered much lower effective RPMs (often under $0.50 per 1,000 views), though TikTok has introduced additional monetization features like Creator Rewards and TikTok Shop that can supplement this. Instagram Reels bonuses vary by invitation and are not consistent enough to use as a baseline for planning.

When comparing platforms, look at your actual earnings data rather than what other creators report. Platform payouts vary so much by niche, geography, and content type that someone else's numbers are rarely a useful proxy for your own.

How to use RPM in your planning

For income planning, use a conservative RPM estimate based on your last 90 days of data (excluding any unusual spikes or dips). If you are new and do not have historical data, research typical RPMs for your niche and platform, then use the lower end of the range.

Build your estimates around three scenarios: a low case (your worst recent month), an expected case (your trailing 90-day average), and a high case (your best recent month). This range-based approach is more honest than picking a single number and projecting it forward indefinitely.