Alert AutoCampaign is 100% FREE to use for a limited number of beta testers! Click here to apply now. ### How do we apply estimated impacts?

Learn how our platform considers the impact of previous optimizations immediately.

### Importance of considering impacts

Why the estimated impact of optimizations should be considered immediately.

Assume a campaign generated \$1 in revenue between 01/01 to 01/03. An optimization event on 01/03 (ie. pausing "Landing Page #3") should increase subsequent revenue by 25%. To not prematurely optimize campaigns then, the estimated impact of this optimization should be applied immediately in calculations. Rather than calculating the bid based on \$1 in actual revenue until 01/03, it should be based on \$1.25 [\$1 actual revenue x 1.(25%) impact].

Similarly, when the next optimization occurs on 01/05, the estimated +10% revenue impact should again be applied to all prior data: \$1.25 "post-events" revenue until 01/03, and \$2 between 01/03 to 01/05. By doing so, the estimated impact of optimizations can be immediately considered in all subsequent optimizations (despite the lagging actual revenue not reflecting the impact yet).  Whether it's the estimated impact of an optimization, payout adjustment, "offline" improvement (such as adding a call center for partials), or the need to discount a freak accident on your data (such as downtime) - it's possible by creating manual "events". Our platform can also continuously calculate the estimated impact of its own automatic optimizations and apply it immediately.

This methodology permits a slew of other benefits - such as monitoring the direction of previous optimizations (by comparing estimates to the actual data), simulating second-price auctions on platforms that don't support it, re-assessing previously paused items, "maximizing" campaign goals rather than simply satisfying them, or even applying "weights" based on the age of data.

#### Optimization methodology

Our methodology can be summed into this formula:
f(c) = Σ [sum(n) x multipliers(n)] | n=0 to ‘c’

Wherein:
- ‘n’ is the event number in the \$events array (starting from 0)
- 'c' is the number of total events
- sum(n) is the raw performance/spend metric total (for every item value) between \$events[n-1][‘timestamp’] and \$events[n][‘timestamp’]
- multipliers(n) is the compounded impact of all events that apply between \$events[n-1][[‘timestamp’] and \$events[n][‘timestamp’]

#### Confused? Let us explain.

Every "event" (ie. optimization) is logged in our system with its estimated impact.

For every event, the raw data is gathered between it ('n') and the previous event ('n'-1). The compounded estimated impact of all subsequent (applicable) events is then applied to this raw data.

Finally, each period's calculated "post-events" data above is summed (Σ). This new total includes the estimated impact of all subsequent events, and therefore, is more accurate than simply optimizing campaigns based on raw lagging data.