When the NAR settlement's practice changes took effect in August 2024, the consensus prediction was simple: decoupling buyer-agent compensation from the MLS would send commissions sliding. Some analysts forecast buyer-side rates falling by a point or more within two years.

Eighteen months of transaction data later, that's not what happened. The story is stranger — and more useful for agents trying to price their services.

The headline findings

• The national average buyer-side commission was 2.42% in Q3 2025 — higher than the 2.36% of Q3 2024, and effectively back at pre-settlement levels (Redfin).
• An analysis of 224,000+ transactions across ~1,300 brokerage offices found rates dipped immediately after the August 2024 changes, then fully recovered: 2.55% by mid-January 2025, identical to a year prior (AccountTECH).
• Agents themselves report total (combined buy + list side) commissions rising from 5.32% to 5.44% in 2025 (Clever's survey of 806 agents).
• The averages hide the real change: bifurcation by price tier. Entry-level homes now pay the highest buyer-side rates since 2023, while luxury commissions compress.
• ShowSmartly's own data agrees from a different angle: across 543 commission disclosures sent through our automated showing replies (December 2025 – June 2026), communicated buyer-side compensation averaged 2.88%, and the median rate on active listings held at 3.0% in every quarter we've tracked. Where compensation is offered at all, it isn't drifting down — it's parked at the traditional number.
• The more interesting internal signal: 44% of new listings entered our system in Q1 2026 with no pre-committed buyer-side offer at all — the number was left open to be negotiated offer-by-offer. By Q2 it had snapped back to 0%. Pre-committing seems to be winning, at least in the Midwest markets we see most.

Method

This analysis draws on four public datasets and one proprietary one: Redfin's quarterly commission reports (transactions handled by Redfin buyer's agents), AccountTECH's analysis of 224,000+ closed transactions from ~1,300 offices, Clever Real Estate's 2025 survey of 806 practicing agents, and a May 2025 Federal Reserve research note on broker compensation trends.

We add ShowSmartly's own anonymized aggregate: buyer-agent compensation as communicated in automated showing responses on our listing-agent accounts — 543 commission disclosures plus 66 listing records with commission terms, December 2025 through June 2026, concentrated in Michigan and other Midwest markets. It's a small, young, geographically skewed dataset and we present it as exactly that. But it measures something the transaction datasets can't: the number listing agents communicate before negotiation, at the showing stage. Zero-value entries (listings with no pre-committed buyer-side offer) are reported separately rather than averaged in. All figures are aggregates; no agent, listing, or client data is identifiable.

No single source is perfect: Redfin's data skews toward its own buyer mix, surveys capture perception alongside fact, and our aggregate reflects ShowSmartly's customer base rather than the full market. Read the convergence, not any single number.

The number that refused to fall

Buyer-side commission averages, national:

Period	Average buyer-agent commission	Source
Q3 2024 (settlement quarter)	2.36%	Redfin
Mid-January 2025	2.55%	AccountTECH (224k transactions)
Q2 2025	back at pre-settlement levels	Redfin
Q3 2025	2.42%	Redfin
Dec 2025 – Jun 2026	2.88% (communicated, pre-negotiation)	ShowSmartly showing-reply aggregate (n=543)

The immediate post-settlement dip was real — AccountTECH's transaction-level data shows it clearly in late 2024. But it behaved like a market absorbing a procedural shock, not a market repricing a service. Within five months, rates had round-tripped.

Why? Redfin's analysts point to a structural reason: in most of the country, 2025–26 has been a buyer's market. Sellers competing for scarce buyers kept paying buyer-side compensation as a deal-making concession — and buyers, holding leverage, had little reason to grind down their own agent's fee when the seller was effectively funding it. The settlement changed where the number is written, not who ultimately bears it. So far. (See Inman's coverage of the rebound.)

The real story: bifurcation by price tier

The flat national average is hiding a scissor pattern (Redfin, Q3 2025):

Price tier	Avg. buyer-side commission, Q3 2025	Direction
Under $500K	2.52%	Up from 2.45% a year earlier — highest since 2023
$500K–$999K	2.32%	Roughly flat (2.34% in Q2 2025)
$1M+	2.22%	Down from 2.24% a year earlier — continued slow compression

Two markets are emerging. At the entry level, the work-per-dollar ratio is brutal — more showings, more financing fragility, more fall-throughs — and percentage-based fees on small denominators produce small checks. Agents are holding or raising rates there, and sellers of starter homes, who need the buyer pool most, are paying them. At the top, dollar amounts are large enough that even modest percentage trims save five figures, and sophisticated parties negotiate accordingly. Luxury is where the settlement's negotiate-everything logic actually bites.

For teams, the operational takeaway: your effective rate is now a portfolio property, not a policy. Knowing your blended rate by price band — and how it compares to these benchmarks — is a KPI conversation. (It's one of the metrics ShowSmartly's team KPI dashboard tracks per agent and per listing band.)

What we see from the showing inbox

ShowSmartly sits at an unusual vantage point: we generate the listing agent's reply to every showing request, which means we see the buyer-side compensation number at the moment it's communicated — before any negotiation, before any offer.

Three patterns stand out in our December 2025 – June 2026 aggregate:

1. Where a number is offered, it's the old number.

The median communicated rate on active listings was 3.0% in Q4 2025, Q1 2026, and Q2 2026 — flat. The average across 543 showing-reply disclosures: 2.88%. Our book is concentrated in Michigan and similar Midwest markets — squarely the sub-$500K tier where Redfin's data shows rates rising — and our pre-negotiation numbers sit right where that tier's closed-transaction averages (2.52%) would predict, plus the negotiation gap you'd expect between the offered number and the closed one.

2. The real experiment is whether to pre-commit at all.

In Q1 2026, 44% of new listings entered our system with no buyer-side compensation set — sellers and listing agents leaving the number open to negotiate offer-by-offer. One quarter later: 0%. In showing-heavy Midwest winters, "we'll discuss it with the offer" appears to have cost more friction than it saved, and the pre-committed number came back.

3. Dispersion is the story everywhere.

The consistent public finding holds in our data too: the MLS used to anchor everyone to the local customary number; without the anchor, the spread — between tiers, between metros, between listings on the same street — is what's widening, even where averages barely move.

(Dataset caveats: 3 brokerage accounts, 66 commission-bearing listings, 543 reply-level disclosures, Midwest-skewed. We'll republish these cuts quarterly as the base grows.)

What 18 months teaches us

1. The MLS field was an anchor, not a price.

Removing it didn't cut prices; it removed the default. Negotiation now happens deal-by-deal — which rewards agents who can articulate value and document everything.

2. Compensation conversations moved into showing workflows.

With offers off the MLS, the disclosure now travels through listing-agent communications — emails, showing replies, flyers. That's a compliance surface (see our audit checklist) and a data surface: whoever handles the showing communications sees offered compensation before anyone else does.

3. Buyer agreements made fees stickier, not looser.

A signed, specific number is psychologically and contractually harder to discount mid-transaction than an MLS default ever was. The settlement may have inadvertently stabilized buyer-side fees for represented buyers.

4. The pressure shows up at the tails.

Luxury compression, flat-fee and hourly experiments, and unrepresented buyers are growing at the edges. The Fed's May 2025 research note frames it as slow structural change with the steepest gradient at high price points — consistent with what the tier data shows.

Predictions for 2027

Made carefully, and we'll grade ourselves in next year's update:

The national average stays in the 2.3–2.5% band absent a major market reversal. The buyer's-market subsidy logic holds while inventory does.
Luxury compression continues — sub-2.1% average buyer-side rates above $1M by late 2027 as flat-fee structures take share at the top.
Entry-tier rates hold above 2.5%, and "seller pays buyer-side as concession" becomes a near-universal feature of sub-$500K listings in soft markets.
Dispersion keeps widening. The interesting chart in 2027 won't be the average — it'll be the interquartile range.
Compensation data becomes a listing-side tool. Teams that track communicated compensation across their market will price listings (and recruit agents) with information their competitors don't have.

Download the full dataset behind this post (CSV)

Both tables — national averages and price-tier breakdown — in one CSV for your own analysis.

We’ll email occasional compliance updates. Unsubscribe anytime.

What 18 Months of NAR Settlement Data Tells Us About Commission Averages

Important Disclaimer