<?xml version="1.0" encoding="UTF-8"?>
<feed xmlns="http://www.w3.org/2005/Atom" xml:lang="en">
  <title>Ed Chen — Field Notes</title>
  <subtitle>Practitioner writing on institutional finance UX and portfolio engineering. FIX 4.4, MiFID II, ASIC RG 268, KYC AML, plus how I run my own portfolio as a product.</subtitle>
  <link rel="self" type="application/atom+xml" href="https://edwson.com/feed.xml"/>
  <link rel="alternate" type="text/html" href="https://edwson.com/notes.html"/>
  <link rel="hub" href="https://edwson.com/notes.html"/>
  <id>https://edwson.com/feed.xml</id>
  <updated>2026-05-11T00:00:00Z</updated>
  <generator uri="https://edwson.com/" version="1.0">edwson.com hand-built Atom feed</generator>
  <icon>https://edwson.com/favicon-32x32.png</icon>
  <logo>https://edwson.com/img/og-image.webp</logo>
  <rights>© 2026 Ed Chen. All rights reserved.</rights>

  <author>
    <name>Ed Chen</name>
    <email>ed@edwson.com</email>
    <uri>https://edwson.com/</uri>
  </author>

  <category term="institutional-finance" label="Institutional Finance"/>
  <category term="fintech-ux" label="Fintech UX"/>
  <category term="regulated-platform-design" label="Regulated Platform Design"/>
  <category term="portfolio-engineering" label="Portfolio Engineering"/>

  <entry>
    <title>Reading AWStats Like a Product Manager</title>
    <link rel="alternate" type="text/html" href="https://edwson.com/note-portfolio-funnel-analysis.html"/>
    <id>https://edwson.com/note-portfolio-funnel-analysis.html</id>
    <published>2026-05-11T00:00:00Z</published>
    <updated>2026-05-11T00:00:00Z</updated>
    <author>
      <name>Ed Chen</name>
      <uri>https://edwson.com/</uri>
    </author>
    <category term="portfolio-engineering" label="Portfolio Engineering"/>
    <category term="awstats-funnel-analysis" label="AWStats Funnel Analysis"/>
    <category term="personal-brand-seo" label="Personal Brand SEO"/>
    <summary type="text">My second-most-visited page had an 84.6% exit rate. The AWStats data showed exactly where the funnel leaked. Three surgical fixes — a first-screen CTA, a sidebar nav link, and a dedicated bridge section — repositioned a 156-entry catalog page as a feeder to four case studies.</summary>
    <content type="html"><![CDATA[
      <p>My design-system-showcase page was the second-most-visited surface on edwson.com — 156 entries over the analytics period — but had an 84.6% exit rate (132 of 156 visitors left from that same page without clicking anywhere else). The page ranks well for "design system + institutional finance" searches, so first-time visitors arrived, scanned the 146-component catalog, and left without ever seeing that those components ship in production across four case studies.</p>
      <p>The fix was not to redesign the catalog. It was to add three insertion points that bridge the page to the case-study layer — a first-screen CTA, a sidebar nav link, and a dedicated "Where this ships" section between the existing System Card and Editorial Stance. Zero regression on the existing 149 sections. Expected funnel lift: ~2× case-study click-through from this entry point.</p>
      <p>Read the full methodology including the falsifiability commitment at <a href="https://edwson.com/note-portfolio-funnel-analysis.html">edwson.com/note-portfolio-funnel-analysis.html</a>.</p>
    ]]></content>
  </entry>

  <entry>
    <title>Why KYC Drop-Off Spikes at EDD</title>
    <link rel="alternate" type="text/html" href="https://edwson.com/note-kyc-drop-off-enhanced-due-diligence.html"/>
    <id>https://edwson.com/note-kyc-drop-off-enhanced-due-diligence.html</id>
    <published>2026-04-18T00:00:00Z</published>
    <updated>2026-04-24T00:00:00Z</updated>
    <author>
      <name>Ed Chen</name>
      <uri>https://edwson.com/</uri>
    </author>
    <category term="kyc-aml" label="KYC AML Onboarding"/>
    <category term="enhanced-due-diligence" label="Enhanced Due Diligence"/>
    <category term="regulated-onboarding" label="Regulated Onboarding"/>
    <summary type="text">The baseline KYC funnel works. Enhanced Due Diligence is where it falls apart. The fix is not to shorten a regulatory requirement — it is to change how the questions are presented. Three interventions lifted EDD completion from 35% to 70%, with FinCEN / FATF / MAS / FCA citations.</summary>
    <content type="html"><![CDATA[
      <p>Standard KYC flows complete at 75-85%. The drop-off cliff lives in Enhanced Due Diligence — where the additional questions imposed on high-risk customers (PEPs, high-volume traders, certain geographies) push the completion rate to 35% in production data I have seen at three institutional firms.</p>
      <p>The fix is not to shorten EDD — that is a regulatory floor under FinCEN 31 CFR 1010.230 and FATF Recommendation 10. The fix is three UX interventions: progressive disclosure with explicit progress, save-and-resume infrastructure backed by magic-link auth, and in-context regulatory citations that explain why each invasive question is being asked.</p>
      <p>Read the full breakdown at <a href="https://edwson.com/note-kyc-drop-off-enhanced-due-diligence.html">edwson.com/note-kyc-drop-off-enhanced-due-diligence.html</a>.</p>
    ]]></content>
  </entry>

  <entry>
    <title>ASIC RG 268 Cross-Border Disclosure — Three UX Patterns</title>
    <link rel="alternate" type="text/html" href="https://edwson.com/note-asic-rg-268-disclosure-patterns.html"/>
    <id>https://edwson.com/note-asic-rg-268-disclosure-patterns.html</id>
    <published>2026-04-15T00:00:00Z</published>
    <updated>2026-04-24T00:00:00Z</updated>
    <author>
      <name>Ed Chen</name>
      <uri>https://edwson.com/</uri>
    </author>
    <category term="asic-rg-268" label="ASIC RG 268"/>
    <category term="cross-border-disclosure" label="Cross-Border Disclosure"/>
    <category term="regulated-trading-ux" label="Regulated Trading UX"/>
    <summary type="text">ASIC RG 268 requires retail clients to see jurisdiction-specific disclosures before each transaction. Three patterns keep recurring across implementations I have seen: modal (rejected), strip (mediocre), and inline contextual (winner). A/B data from three production studies.</summary>
    <content type="html"><![CDATA[
      <p>ASIC Regulatory Guide 268 requires retail clients of an Australian Financial Services Licence (AFSL) holder to see jurisdiction-specific risk disclosures before each transaction. The "meaningful engagement" test is what determines whether the disclosure satisfies the regulator.</p>
      <p>Three patterns recur across implementations: modal blocker (rejected by users and by ASIC as "click-through ritual"), persistent strip (mediocre because users habituate), and inline contextual disclosure adjacent to the trade input (winner in three A/B studies). The inline pattern shipped at ACY (AFSL 403863) survived multiple ASIC review cycles without triggering a single material finding.</p>
      <p>Read the pattern comparison with citations at <a href="https://edwson.com/note-asic-rg-268-disclosure-patterns.html">edwson.com/note-asic-rg-268-disclosure-patterns.html</a>.</p>
    ]]></content>
  </entry>

  <entry>
    <title>Designing MiFID II Best-Execution Reports for Buy-Side PMs</title>
    <link rel="alternate" type="text/html" href="https://edwson.com/note-mifid-ii-best-execution-ux.html"/>
    <id>https://edwson.com/note-mifid-ii-best-execution-ux.html</id>
    <published>2026-04-08T00:00:00Z</published>
    <updated>2026-04-24T00:00:00Z</updated>
    <author>
      <name>Ed Chen</name>
      <uri>https://edwson.com/</uri>
    </author>
    <category term="mifid-ii" label="MiFID II"/>
    <category term="best-execution" label="Best Execution"/>
    <category term="buy-side-ux" label="Buy-Side UX"/>
    <summary type="text">MiFID II Article 27 mandates venue-quality reports (RTS 27) and top-five-venue reports (RTS 28). The design answer is three layers: a policy-vs-observed compliance strip, a filterable venue attribution grid with slippage in basis points, and a trade-level audit drawer with CSV export — replacing the unreadable PDF most firms generate.</summary>
    <content type="html"><![CDATA[
      <p>MiFID II Article 27 forces firms to prove best execution. The default output is a quarterly PDF nobody reads. The data underlying that PDF — fill rates, slippage by venue, time-to-fill distributions — is exactly what a portfolio manager needs to improve execution quality, but it never reaches them in usable form.</p>
      <p>The design fix is a three-layer screen: a policy-versus-observed compliance strip at the top (compliance officer scan), a filterable venue attribution grid in the middle (the PM's primary workspace), and a trade-level audit drawer on demand (TCA team workflow). Slippage in basis points, not currency. Time-to-fill as a distribution, not a mean.</p>
      <p>Read the full three-layer breakdown at <a href="https://edwson.com/note-mifid-ii-best-execution-ux.html">edwson.com/note-mifid-ii-best-execution-ux.html</a>.</p>
    ]]></content>
  </entry>

  <entry>
    <title>Why FIX 4.4 Latency Dictates Order Entry Form Design</title>
    <link rel="alternate" type="text/html" href="https://edwson.com/note-fix-latency-order-entry-design.html"/>
    <id>https://edwson.com/note-fix-latency-order-entry-design.html</id>
    <published>2026-04-01T00:00:00Z</published>
    <updated>2026-04-24T00:00:00Z</updated>
    <author>
      <name>Ed Chen</name>
      <uri>https://edwson.com/</uri>
    </author>
    <category term="fix-4-4" label="FIX 4.4 Protocol"/>
    <category term="institutional-trading" label="Institutional Trading"/>
    <category term="order-entry-ux" label="Order Entry UX"/>
    <summary type="text">When an order ticket has a 10ms round-trip budget, the whole form-design playbook from consumer apps goes in the bin. Pre-flight validation, reversible execution, keyboard-first shortcuts, and a normalized FIX reject taxonomy. Four years inside that disorientation at ACY Securities.</summary>
    <content type="html"><![CDATA[
      <p>FIX 4.4 round-trip latency to a tier-1 venue is 8–12ms. Consumer-app safety patterns — "Are you sure?" confirmations, slow async state transitions, no keyboard shortcuts — are a latency tax that breaks institutional workflow. Trader productivity drops to a fraction of what the system can support.</p>
      <p>Four design changes shipped at ACY Securities: pre-flight validation that runs before the FIX message is built (no round-trip on validation), reversible execution UI that allows post-fill cancellation when the venue protocol supports it (FIX 35=F), keyboard shortcuts that map to FIX tags, and a normalized reject-code taxonomy that reduced support calls by ~70% by translating venue-cryptic reject codes into actionable trader language.</p>
      <p>Read the full design playbook at <a href="https://edwson.com/note-fix-latency-order-entry-design.html">edwson.com/note-fix-latency-order-entry-design.html</a>.</p>
    ]]></content>
  </entry>

</feed>
