In nearly 20 years in search marketing, something I continue to see and debate about is shared budgets.
When I ask the advertiser why, I always get a similar answer: “We have a set limited budget, and this is the best way for us to fund everything.”
Or, the one that truly pains me: “The engine sent me a rep for the fiscal quarter who recommended it along with a portfolio bid strategy to let it all work together.”
That is a response I get from mom-and-pop operations and SMBs, all the way up to major holding company ad agencies and Fortune 100 enterprise brands (if you fall into the latter and you went with the platform recommendation without questioning it, shame on you).
The reality is that sometimes this approach works fine, and other times, you are inadvertently shooting yourself in the foot, as your shared spend ends up limiting your potential return.
The Situation
A shared budget (often called a portfolio budget) does, in fact, help bandwidth-strapped operations (i.e., literally every marketing team on Earth) avoid overspending in a single engine. It can also complement a portfolio bid strategy in some scenarios positively.
Image from author, October 2024
However, a shared budget rewards the high-volume or high-demand elements first.
It’s essentially like attempting to work in the front office of the NY Jets – those who come in fast and with deep pockets get big-name players (who often don’t pan out), and everyone else loses out (especially the fans).
The Impact
Pairing a shared budget is cute until you realize you have a low conversion rate-producing campaign that has a high volume in traffic, sharing a budget with a campaign that has a low demand in traffic but is producing a high conversion rate.
Or, god forbid, you share a budget across campaigns that have different bid strategies or across different channels such as video and search (if you are mentally saying you’re doing this now, then you have messed up #ChangeMyMind).
Thus, the underproducing but high-volume campaign will take a disproportionate amount of the budget based on daily demand, potentially diminishing the opportunity or amount of time the high-producing campaign can show (as viewed in Impression Share and Impression Lost to Budget).
This is referred to as internal budget cannibalization.
Or when one entity takes a disproportionate amount of the total allotted budget against the entities with which it shares the funds.
Image from author, October 2024
This, in turn, reduces the aggregate performance (which can only be looked at as aggregate because they share the budget) of the campaigns sharing the budget.
To put this in a different perspective: This is like claiming all New York NFL teams are the best because the Jets and Giants combined have won five championships, but the Chiefs have only won four.
NY shares an NFL affiliation and has historical positive performance, but that can override recent positive performance of the Chiefs (assuming these two things were actually search campaigns sharing a budget and not pro football teams).
The Fallout
If you’ve read any of the above, this should be fairly straightforward, but to lay it out nicely: You treated two campaigns like they were equals and told them to share some money evenly.
This potentially will have left you with a missed opportunity for traffic and conversions.
If you share a budget between campaigns that have different themes (brand vs. non-brand), channel goals (search vs. video), or objectives (traffic vs. conversions), the problems will be immediately amplified.
The Fix
I should note that every operation is different. Some may not fall into this scenario, and some will – I see it more often in smaller agencies and brands strapped for bandwidth.
My personal preference, and I exercise this approach whenever I can, is to have standalone, daily budget caps for the individual campaign.
Note: I said daily. If this campaign is ongoing or evergreen, don’t use a campaign total; it’ll become a pain in the butt for you later. Only use campaign total for short campaign flights with predefined end dates.