The Highest Volume Bid Strategy is the default. You’re using that strategy without realizing it. The algorithm will try to get you the most optimized actions within your budget.
But, you also have the option of using a manual bid strategy, which is usually a Cost Per Result Goal or Bid Cap.
Should you be using them? Or are there specific times when you should?
In most cases, just stick with the default. Since Meta is already trying to get you the most optimized actions within your budget, overriding that strategy will rarely be beneficial. But there are exceptions.
You might consider manual bidding when your performance goal is to maximize conversions, especially when that conversion is a purchase, or maximize value. If there’s a specific cost per purchase that you can’t go over to remain profitable, it can be a good option.
But it won’t create miracles as it will most likely impact volume, so make sure that you’re spending enough or your ads may stop delivering — or underdeliver. You may also run into optimization and learning phase issues if you are already on the edge of 50 conversions per week.
I would not advise using manual bidding when using a performance goal for any top-of-funnel action (link clicks, landing page views, post engagement, ThruPlay, etc.) or possibly even leads.
By restricting the algorithm, the focus will be on low-cost actions, so Meta will go to specific sources to find those actions at a lower cost. This will almost always be at the expense of quality.
Of course, top-of-funnel optimization always has the potential to be problematic anyway concerning quality, but manual bidding can make it worse. And if you are struggling to generate quality leads, setting a Cost Per Result Goal is unlikely to help.
This is one of those cases where you’re often better off just letting Meta handle it automatically, beyond the rare exception when your optimized action is a purchase. If you’re looking for a magical way to generate cheap engagement or leads, this is unlikely to be a solution — other than generating surface-level results.
In those cases, you’re not going to outsmart the algorithm, so don’t even bother.