Adviser Directory
How We Rate Financial Advisers
Every figure on a firm’s page comes from a public record and links to it; on the hand-decoded pages, each was checked by a program and a second reviewer before publication.
Which firms are included
Every SEC-registered investment adviser headquartered in the 13 western states that serves at least 100 individual clients has a profile: 1,729 firms. All of them are scored with the same eight factors and formulas; what differs is how much of each firm’s record has been read, and each page names its tier. Firms cannot pay to be listed or rated, and no firm reviewed its page before publication.
How the tiers differ
- Hand-decoded (149 firms: 102 in Washington, Oregon and Idaho that offer financial planning and portfolio management to individuals and serve at least 500 individual clients, and 47 of the largest advisers across the 13 western states, decoded in November 2026): every factor read by hand from the full filings, with an outside-record search of the SEC, FINRA and the home-state securities regulator. As of Nov. 13, 2026, 148 of these ratings are published and 1 is held back.
- Rated from filings and federal records (230 firms): every other adviser with at least five scored factors (regulatory record, complaint record, at least 10 client reviews, cost and fee transparency) is rated from its filings and federal records, with confidence no higher than Medium. The fee schedule comes from the firm’s Part 2A brochure, read twice by independent readers; the regulatory record comes from Form ADV, FINRA BrokerCheck and a reviewed name search of SEC enforcement records.
- Profile, not rated (1,350 firms; 1,341 with the brochure read for its fees): a rating needs at least 5 of 8 factors scored and these firms have at most 4. Fee transparency and, where a fee on $1 million can be worked out, Cost are scored from the brochure; the brochure readings of conflicts and fiduciary standing failed an independent blind check and are not used, disclosure consistency is not collected, and client reviews count only with at least 10 public reviews. The profile shows the firm’s Form ADV answers, disclosure pages, fee schedule and BrokerCheck record. How these profiles work.
Where the data comes from
For each of the 149 hand-decoded firms: its Form ADV Part 1, Form ADV Part 2A brochures and Form CRS, downloaded from the SEC’s Investment Adviser Public Disclosure site; the SEC’s monthly adviser data, including one file a year for the firm’s history; that site’s records of the firm’s current advisers, in aggregate; Google review totals for the listings tied to the website the firm reports; and Better Business Bureau listings matched to the firm’s filed address or phone. The 20 of them serving the most individual clients were also researched in depth: regulator, FINRA and court records, other review sites, news coverage and the firms’ own marketing; the other hand-decoded firms whose ratings could not otherwise be published, and the 47 firms added in November 2026, had their SEC, FINRA BrokerCheck and home-state regulator records searched, and those searches were re-checked by a reviewer who did not run them. Every other West adviser is profiled from its filings and federal records, as described in the tiers above.
How every fact is checked
Every fact is stored with the exact quote it rests on and the page it came from. A program confirms each quote appears in the cited document. Then a separate reviewer re-reads the documents, including fee tables printed as images, and corrects anything the record doesn’t support. Close calls are settled by written rules applied the same way to every firm, for example: a firm whose advisers also sell commissioned products is described as a fiduciary only when giving investment advice. The adviser scoring program was written by the research team and later amended for versions 2.5 and 2.6. Its copy in a second program was written by the same people, so it checks the arithmetic, not the reading of the rules (149 firms, 33 checks each, no differences). The independent check is a re-implementation written by a separate reviewer from the published rules alone, without reading either program: on Nov. 13, 2026 it re-scored all 121 hand-decoded advisers under version 2.7 and matched every factor score, overall score, confidence level and publication decision.
How reviews and complaints are searched
Client reviews are searched for every adviser: the reviews counted are those on Google business listings matched by the firm’s official website and its Form ADV offices. The Better Business Bureau is searched too, and its listings supply the complaint record. 374 advisers have at least 10 counted reviews and are scored; a firm with fewer than 10 is not scored. The Complaint record follows scoring version 2.7: a complete recorded search of the Better Business Bureau (legal, former and trade names, with city and state) and, for a FINRA member, the arbitration awards in its BrokerCheck report, that finds nothing scores 10. A possible listing that cannot be matched to the firm or ruled out leaves the factor not scored. 1,728 advisers have a Complaint record score. Official websites were checked and corrected for every adviser. 1,572 adviser profiles show the firm’s own logo.
How fee schedules are read
The fee schedule in each adviser’s Form ADV Part 2A brochure is read by two independent readers (1,718 advisers). Every disagreement is settled from the source page, and every figure is stored with its quote and its page. The readings give 1,202 advisers a Cost score and 1,697 a Fee transparency score. Two written rules keep every reading the same.
- Rule D2, tier method: a schedule is blended (each band is charged at its own rate) or whole-account (one rate applies to the whole amount); the brochure’s worked example decides first, then its wording, and a schedule the brochure does not settle is marked unknown and gets no dollar figure.
- Rule D14, primary schedule: when a brochure prices several asset-based schedules, the primary schedule is the one the brochure calls standard or typical, otherwise the schedule of the firm’s main portfolio management service for individual clients when exactly one fits; if more than one fits equally, there is no primary schedule and Cost is not scored.
An independent blind re-read of a random sample matched 96.7% of fee-schedule fields (772 of 798). The sample covered 448 fields of firms read in this round (431 matched, 96.2%) and 350 fields of firms published earlier (341 matched, 97.4%). Every fee fact on a firm's page shows its quote and its page.
How the rating is computed
Each of the eight factors is scored from zero to 10 by a fixed formula; the rating is the weighted average on a 100-point scale, rounded once at the end. Three factors rest on records the firm does not write (client reviews, its regulatory record and its complaint record); the other five come from its own filings. The median rate on $1 million used for Cost is 1% a year, across every firm in the directory with a computable schedule.
| Factor | Weight | What it measures |
|---|---|---|
| Client reviews | 20% | Star ratings and review counts on public platforms. Every adviser is searched: the reviews counted are those on Google business listings matched by the firm's official website and its Form ADV office addresses; the Better Business Bureau, Trustpilot and Facebook are also read for the firms researched in depth. A firm with fewer than 10 counted reviews is not scored; 374 advisers have 10 or more. A firm with fewer than 25 takes a small deduction. Directories and lead-generation sites never count. |
| Regulatory record | 15% | Resolved actions by the SEC, FINRA or state regulators against the firm or a firm it controls; recent actions and large sanctions weigh more. Scored from a completed regulator search, or from the firm's Form ADV Item 11 when every answer there is No or each reported event has been traced and FINRA's records searched. Pending matters are shown but not scored, and individual advisers' records are not part of a firm's rating. |
| Complaint record | 10% | Better Business Bureau complaints in the past three years relative to the firm's number of clients (counted only from a listing matched to the firm's filed address or phone), and customer arbitration awards against the firm. Since version 2.7 a complete recorded search (the Better Business Bureau under the firm's legal, former and trade names with city and state, and FINRA BrokerCheck arbitration awards for a firm that is or controls a broker-dealer) that finds nothing scores 10. A search that was not completed, or a possible listing that cannot be matched to the firm or ruled out, leaves the factor not scored. 1,728 advisers have a Complaint record score. |
| Cost | 15% | What the firm's primary fee schedule costs on a $1 million account compared with the median Northwest adviser. The schedule is read from the firm's Form ADV Part 2A brochure. Not scored when no primary schedule or no rate on $1 million can be worked out; 1,202 advisers have a Cost score. |
| Fee transparency | 10% | How completely the firm publishes its fees: a full breakpoint table that says how the tiers apply scores highest; a table without that, a range, a maximum or no rate at all scores lower. 1,697 advisers have a Fee transparency score. |
| Conflicts of interest | 15% | The conflicts of interest the firm discloses, such as securities commissions, insurance sales or revenue sharing, weighted by how directly they can affect advice. Insurance commissions count once, as insurance sales; benefits from custodians are shown but not scored because nearly every firm has them. |
| Disclosure consistency | 10% | Whether the firm's own client documents (brochure, Form CRS and Form ADV) contradict each other on fees, minimums, what a fee covers, conflicts or disciplinary history, and whether its Form CRS is less than two years old. Only the documents count, not the firm's marketing. |
| Fiduciary standing | 5% | Whether the firm acts as a fiduciary at all times, or only when giving investment advice because its people also sell commissioned products. |
A factor with no usable data is left out and the remaining weights are scaled up; missing data never counts for or against a firm. Each firm’s page lists every point deducted, with its source, and every factor that could not be scored.
Documents, not marketing
Disclosure consistency counts only contradictions between the firm’s own client documents: two statements in its brochure, Form CRS or Form ADV that cannot both be true, each quoted with its page. A mismatch with a Form ADV checkbox alone is shown on the firm’s page as a filing inconsistency but not counted, and neither is a document that is silent or wording that can be read both ways. Each contradiction costs 2 points, up to 6, and a Form CRS more than two years old costs 2.
Where a firm’s website says something its filings contradict, for example that it sells no products while its brochure says its advisers earn insurance commissions, the firm’s page quotes both. Those findings are shown but not scored: only the 20 firms researched in depth were checked for marketing claims, and a deduction that only researched firms could receive would measure the research, not the firms.
Confidence
Each rating carries a confidence level built from points: one for each of five groups that could be scored (client reviews; the regulatory record; the complaint record; cost, fee transparency or conflicts; disclosure consistency and fiduciary standing), one for 15 or more distinct sources, one when both the in-depth research and the reading of the filings were independently verified, one when the outside-record search was re-checked by a reviewer who did not run it, and one more when client reviews are scored, less one if the reviewer had to correct more than 30% of the facts. Six or more points is high, four or five medium, three or fewer low. A rating from filings and federal records is capped at medium.
Measured accuracy
Three blind checks measured how accurate the hand-decoded ratings are. In each, readers who had not seen the decoded data re-read the filings of 12 firms picked at random, and a third reader settled every difference against the filing text. In the first sample, 4.1% of the fields read were wrong, and 10 of the 12 overall scores were within 3 points of the corrected value; those errors were corrected, and the judgment calls for the original 102 firms were read again. In the second sample, of 12 different firms, 2.8% of fields were wrong after that review, and all 12 overall scores were within 3 points of the corrected value. The third sample, of 12 of the 50 large western advisers decoded in November 2026, found 3.5% of fields wrong after review, and 11 of the 12 overall scores were within 3 points of the corrected value; 19 of those 50 advisers were then published. The errors found were corrected. Because each sample is above the 2% limit for individual facts, hand-decoded ratings are shown at no more than medium confidence, with a margin of about plus or minus 3 points on the overall score.
The fee schedules behind the ratings of advisers outside the hand-decoded set were checked separately; the result is in the section on how fee schedules are read.
When a rating is published
A rating is published only when all four of these hold:
- At least five of the eight factors are scored.
- At least one factor rests on a record outside the firm’s own filings: client reviews (10 or more), the complaint record, or a regulatory record from a completed search of the SEC, FINRA and the securities regulator of the firm’s home state (for a rating from filings and federal records: the reviewed SEC enforcement search and the FINRA BrokerCheck record). A regulatory record scored only from the firm’s own Form ADV answers still counts in the score but not toward this test: a firm’s own filings alone never make a rating.
- Confidence is medium or high.
- At least 60% of the rubric’s weight is scored (version 2.5).
One rating is held because confidence is low. A held firm's page shows no number; it says why the rating is held and shows the same records as every other firm. Of the 1,350 advisers profiled without a rating, 1,341 had their latest Part 2A brochure read for its fees, twice by independent readers with differences settled from the source page: 1,197 of those profiles show a fee schedule and 858 show a fee on a $1 million account. The fee readings give a Fee transparency score to 1,318 of them and a Cost score to 858, but a rating needs at least 5 of the 8 factors scored and none has more than 4 of the 8, so none is rated. Client reviews are scored only for a firm with at least 10 public reviews, disclosure consistency is not collected, and conflicts of interest and fiduciary standing are not scored: a blind check of the brochure readings found 20 differences in 228 fields, or 8.77%, above the 1% limit for a published fact. 230 other advisers reached five scored factors and are rated from filings and federal records, at Medium confidence at most. These are the adviser-only profiles: counting the 1 hand-decoded adviser whose rating is held, 1,351 advisers have no published rating.
How are firms ranked, and how sensitive are the scores?
Every rating in the underlying records (and its record-tier mirror) carries eligibility for ranking, ranking note, percentile within its type and rank within its type, the same fields as every other firm type: a rating is rank-eligible when it is published, at least 85% of the rubric's weight is scored, and it does not reach the top tenth only because a factor could not be scored. Percentiles and ranks are over rank-eligible firms only. Because client reviews (20%) are rarely scorable, 25 of 120 published ratings are rank-eligible.
Weight sensitivity (120 published ratings; each weight moved 5 points, others unchanged, renormalised per firm; overall from the stored one-decimal factor scores):
| Factor | Shift | Spearman vs published | Median change | Max change | Firms moving 3+ | Top 10% kept |
|---|---|---|---|---|---|---|
| reviews | +5 | 0.9995 | 0 | 1 | 0 | 100% |
| reviews | -5 | 0.9994 | 0 | 2 | 0 | 100% |
| regulatory | +5 | 0.9983 | 1 | 4 | 4 | 93% |
| regulatory | -5 | 0.9982 | 1 | 4 | 6 | 100% |
| complaints | +5 | 0.9988 | 1 | 4 | 6 | 93% |
| complaints | -5 | 0.9994 | 1 | 4 | 9 | 100% |
| cost | +5 | 0.9983 | 0 | 3 | 1 | 93% |
| cost | -5 | 0.9974 | 0 | 3 | 1 | 100% |
| transparency | +5 | 0.9937 | 1 | 3 | 8 | 93% |
| transparency | -5 | 0.9917 | 1 | 4 | 14 | 100% |
| conflicts | +5 | 0.9957 | 2 | 5 | 33 | 86% |
| conflicts | -5 | 0.9939 | 2 | 5 | 34 | 100% |
| disclosure | +5 | 0.9972 | 1 | 4 | 7 | 100% |
| disclosure | -5 | 0.9958 | 1 | 5 | 10 | 100% |
| fiduciary | +5 | 0.9970 | 1 | 3 | 1 | 86% |
| fiduciary | -5 | 0.9942 | 1 | 3 | 4 | 79% |
Data-gap test: each N/A factor scored at the median of the published firms that have it instead of renormalising. 103 published ratings have a gap; rank correlation 0.9919, 50 firms move 3 points or more (largest 24). Firms in the top tenth only because of a data gap: CRD 108917 (95, 94 with medians); CRD 310737 (95, 94 with medians); they are not rank-eligible.
Plain finding. The data-gap test moves 50 of the 103 published ratings that have an unscored factor by 3 points or more (largest 24). The "about ±3 points" margin on each rating covers decoding errors measured in the blind samples only; it does not cover how missing factors are treated. Each published rating now carries change from a data gap and a margin note that states both; the 3 rank-eligible firms whose overall moves 3 or more points under the test (of 25 rank-eligible) carry ranking caution. The rubric is not changed.
What changed in each version?
- Version 2.7, Nov. 12, 2026. A complete recorded search (BBB by legal, former and dba names with city and state; FINRA BrokerCheck arbitration for a FINRA member) that finds nothing scores 10; found complaints and awards keep the existing deductions; never searched, or an unresolved possible BBB listing, is N/A. Applied to the 121 hand-decoded advisers (then the rating program) and to the record tier.
- Version 2.6, Nov. 10, 2026: one rounding rule for every firm type; ranking fields. Factor scores are now rounded half-up to one decimal and the overall is computed from them (it had used unrounded factors, so 3 published overalls did not recompute from the scores shown).
- Version 2.5 data update, Nov. 10, 2026: the version 2.4 trial. 19 trial firms published as hand-decoded ratings: 89 published to 108.
- Version 2.5, Nov. 10, 2026: coverage rule shared by every firm type; no factor formula or weight changed.
- Version 2.4, Nov. 10, 2026: decoding rules and a re-read of every hand decode. No factor formula or weight changed. Result: 54 overalls moved and 18 changed band. All 100 published firms stay published.
- Version 2.3, Nov. 10, 2026: brochure-tier trial result.
- Version 2.3, Nov. 10, 2026: tiers; no factor formula or weight changed.
- Version 2.2, Nov. 9, 2026: evidence sources; no weight or factor rule changed. Applied to the 68 held firms: 67 searches complete (SEC IAPD record + DRPs, FINRA BrokerCheck firm record and name searches, home-state regulator lists. Result: 59 published (was 34), 43 held.
- Version 2.1, Oct. 14, 2026: rounding.
- Version 2.1, Oct. 14, 2026. Commissions means securities or investment-product commissions.
- Version 2.1, Oct. 14, 2026: regulator search scope.
- Version 2.1, Oct. 14, 2026. What changed: Marketing-claim contradictions no longer deduct; publication gate.
- Version 2.0, Oct. 14, 2026. Five factors split into eight.
- Version 1.2, Oct. 14, 2026. Regulatory record also usable when every Item 11 'Yes' is traced to its DRP and classified firm vs. individual and a FINRA disciplinary-actions search was completed.
- Version 1.1, Oct. 14, 2026. Regulatory record usable from a Form ADV Item 11 with every answer No. Publication floor: at least 3 of the 5 factors scored, otherwise Low and held back.
- Version 1.0, Oct. 14, 2026.
Scores are comparable within a firm type, not across types: each type is scored on its own rubric from different records, so a median of 97 for hedge fund managers and 73 for credit unions does not mean one kind of firm is safer or better than the other. See the ratings, the fee report or the full directory. Scoring version 2.7; SEC data as of Nov. 2, 2026 for 147 firms, last read Nov. 11 to 13, 2026; Oct. 2, 2026 for 2, last read Oct. 14, 2026. This is information, not investment advice; a firm’s agreement with a client governs its actual fees and services.