So I typically get a group of prospective marketers and ask each one how they'd spend (eg. $1000), de-duplicate and then do what they suggest. I give the business to the one who's campaign gave the best results.
This doesn't guarantee you will get the marketer that will get you the best results. It does guarantee you will find the one with the best intuition for the next step in your marketing process improvement. Rinse repeat every few months.
As an aside, for DIYers, I think Google is back on top. Their AI efforts are having a huge impact on their ability to surface relevant ads. A DIYer can approximate what the big brands are doing with Smart Goals and Automated Insights.
The process I described above is kind of like a single step in an ai type optimization process to find the global maximum of your marketing spend.
I'd be inclined to recommend someone part-time/contract if you have a reasonably narrow or marketing skillets required. Mainly for 1) You'll probably get better rates going direct vs a company that has to layer a margin on-top. 2) If your company grows its easier to bring in a person contributing to your success than if they are contracted to a larger company.
If you need broad marketing skill-sets, then a agency may be better. They can co-ordinate work across the many people within their agency. If you do go a firm/agency I would really recommend being careful you dont get their talent pitching you, then the work is handed to the cheap drone. This is a problem in many agencies.
Most importantly, do you know what marketing work you need?