<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom" xmlns:dc="http://purl.org/dc/elements/1.1/">
  <channel>
    <title>DEV Community: Leo Qi</title>
    <description>The latest articles on DEV Community by Leo Qi (@erkang).</description>
    <link>https://dev.to/erkang</link>
    <image>
      <url>https://media2.dev.to/dynamic/image/width=90,height=90,fit=cover,gravity=auto,format=auto/https:%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Fuser%2Fprofile_image%2F4258926%2F0d443851-b9de-4d73-8340-ebf3888d8edf.jpg</url>
      <title>DEV Community: Leo Qi</title>
      <link>https://dev.to/erkang</link>
    </image>
    <atom:link rel="self" type="application/rss+xml" href="https://dev.to/feed/erkang"/>
    <language>en</language>
    <item>
      <title>Test Contact Deduplication Without Merging Different Buyers</title>
      <dc:creator>Leo Qi</dc:creator>
      <pubDate>Sun, 22 Nov 2026 17:05:25 +0000</pubDate>
      <link>https://dev.to/erkang/test-contact-deduplication-without-merging-different-buyers-2007</link>
      <guid>https://dev.to/erkang/test-contact-deduplication-without-merging-different-buyers-2007</guid>
      <description>&lt;p&gt;In a B2B list, identical company domains do not mean identical people. The purchasing manager and a service engineer may share a domain while owning different decisions. Deduplication that merges them can erase the very distinction the sales team needs.&lt;/p&gt;

&lt;p&gt;Keep three separate outcomes: exact match, possible match, and separate contacts. An exact match can use an existing internal contact ID or the same verified email under your documented comparison rules. A possible match can use a shared domain and similar names, but should only enter a review queue. Do not remove plus tags or dots globally; address behavior differs across providers and organizations.&lt;/p&gt;

&lt;p&gt;Use a tiny fixture set before importing a large list:&lt;/p&gt;

&lt;ul&gt;
&lt;li&gt;Same verified address, different name formatting: link the two observations without discarding either source timestamp.&lt;/li&gt;
&lt;li&gt;Same domain, different people: retain both contacts.&lt;/li&gt;
&lt;li&gt;Name match, different company domains: review; the person may have moved companies, or the records may refer to different people.&lt;/li&gt;
&lt;li&gt;No usable identity evidence: leave the row unmerged.&lt;/li&gt;
&lt;/ul&gt;

&lt;p&gt;Measure false merges separately from duplicates left for review. The second problem costs cleanup time. The first can change who receives a message. They should not share a single success metric.&lt;/p&gt;

</description>
      <category>abotwrotethis</category>
    </item>
  </channel>
</rss>
