Size does not predict quality. That is measured, not assumed: in IKAg's Agency Barometer the rank correlation between headcount and growth is 0.003, which is zero. The median growing agency had 7.1 full-time equivalents, the median shrinking one 7.8.
What size does predict are three very practical things.
1. Whether you get the people you met in the pitch
At a small agency the people presenting are the people working. At a large one it depends on your budget relative to their other clients: if you are their tenth largest, you get their tenth team. So always ask for names, and put them in writing at the start.
2. What happens when someone leaves
An agency of eight that loses two is a different agency. An agency of eighty barely notices. With median staff turnover of 34.6 percent a year in this market, that is not a theoretical risk.
3. Whether they can absorb your peak and your procedures
Large advertisers with a purchasing department impose requirements on insurance, data security and payment terms that an agency of five cannot carry. Conversely, a small budget at a large agency rarely gets attention, however warm the meeting was.
The rule of thumb that works
You want to be big enough to matter at your agency, and small enough to get its best people. If your assignment sits among that agency's five largest clients, you are well placed. If you are the thirtieth, you are buying a name and a junior.
The dip in the middle
From the same measurements: agencies between roughly 15 and 40 full-time equivalents carry the lowest margins in the market, and above 40 the EBITDA margin recovers to 17.1 percent. That is no reason to avoid that group, but it is a reason to ask about their plans: it is the group under most pressure to grow or merge.
See also: how big the average agency is and who works at an agency.