The peer group is the load-bearing wall of executive compensation disclosure. Every percentile you cite, every “competitive positioning” claim, every defense of a large grant rests on it. Get it right and the rest of the conversation is arithmetic. Get it wrong and nothing downstream can be salvaged, because every number you publish is measured against a comparison set that a reader has already decided not to trust.
Here is what I see fail most often at small caps, in rough order of frequency.
1. Peers that are too big
This is the most common finding and the easiest to avoid. A committee assembles a group of companies it admires, or competes with for talent, or hopes to resemble in three years. Half of them are five to twenty times larger.
The problem is mechanical rather than moral. Executive pay scales with company size, so a group skewed larger drags every percentile down. Your CEO’s package lands at the 40th percentile of a peer group whose median company is eight times your revenue — which quietly means the package is well above the market for a company your actual size, and the disclosure is now doing the opposite of what it was meant to do.
The workable rule for most small caps: peers within roughly one-half to two times your size on your most relevant measure, with the group’s median landing near your own value. Aspirational peers are a legitimate secondary reference for talent-market context. They are not a defensible primary benchmarking set.
2. Too few peers
Below roughly a dozen companies, a peer group stops being a distribution and becomes a small collection of anecdotes. One company’s outlier CEO package moves your median materially. Worse, a thin group invites the obvious question: out of thousands of public companies, why exactly these eight?
Small caps sometimes protest that their niche has no real comparables. Occasionally that is true. Far more often it means the screen was drawn too narrowly — a four-digit industry code when a broader classification plus a size band would have produced a defensible group of eighteen.
3. A group that changes without explanation
Peer groups should evolve. Companies get acquired, business models change, you outgrow the set. Turnover is expected.
Unexplained turnover is not. If four peers left and four arrived and the proxy says nothing about it, a reader with a prior year’s document — and both proxy advisors and any serious institutional holder have one — can compare the lists in about thirty seconds. Silent replacement of lower-paying peers with higher-paying ones is one of the few things in this discipline that looks bad even when the underlying reason was perfectly innocent.
The fix costs one sentence: “Three peers were removed following acquisition and two were added to reflect our expansion into diagnostics.”
4. Industry classification drift
A company describes itself one way in its investor deck, is classified another way by its SIC code, and lands in a third bucket in the classification proxy advisors actually use. All three can be defensible; the trouble comes when nobody notices they diverge until an advisor’s independently constructed peer group looks nothing like yours.
You cannot control how third parties classify you. You can control whether you know about the mismatch before proxy season, and whether the disclosure explains the business-model reasoning behind your own group. A company that says plainly why it benchmarks against medical-device companies rather than the broader healthcare bucket has answered the question in advance.
5. No disclosed selection criteria
The most consequential failure, and the most common of all. The proxy lists twenty peers and never says how they were chosen.
A list of names is not a methodology. It cannot be checked, replicated, or defended, and it is indistinguishable from a list assembled to produce a desired answer. What a reader needs is short and structural: the industry definition, the size range, the number of companies that met it, and any judgment calls you made on top.
Three sentences will usually do it:
Peers were selected from GICS 3520 (Pharmaceuticals, Biotechnology & Life Sciences) with market capitalizations between $75 million and $600 million as of June 30. Eighteen of twenty-three companies meeting these criteria were included. Five were excluded as pre-revenue with no named executive officer compensation disclosed.
That paragraph does more for a contested compensation decision than another twenty pages of narrative.
What the filings actually show
Our weekly sweep of SEC filings reads every Item 5.02 8-K — the disclosure a company files when it appoints an officer. In one recent week, 173 companies filed one. Thirteen were genuine CEO or CFO appointments at companies in the $20 million to $1 billion range, with compensation terms disclosed.
Not one of the thirteen named a peer group.
That is not a criticism of the 8-K, which is not where a peer group belongs. It is an observation about sequence. The package is announced first and benchmarked later, if at all — which means the number the market sees on day one was set against something informal, and the committee then spends the following spring constructing a defense of a decision it has already disclosed.
Doing it in the other order
The peer group is cheap to build and expensive to retrofit. It is public information: proxies, 8-Ks and 10-Ks, all of it on EDGAR, all of it citable to a specific accession number.
RepCor builds that group from filings — screened on disclosed criteria, size-appropriate, with percentile placement for base, bonus and total direct compensation, pay mix, and governance flags, every figure tied to its source filing. One business day. $1,500 for the Core report; $2,500 for the Production tier, which adds overhang and pay-versus-performance.
It is the cheapest input to a decision you will be disclosing, and defending, for years.
Robert Steele is the CEO of RepCor and a sitting small-cap public-company CFO. Nothing here is legal advice.