diff --git a/hugolib/page.go b/hugolib/page.go
index b5d13d4db2c..5de976bd492 100644
--- a/hugolib/page.go
+++ b/hugolib/page.go
@@ -18,6 +18,7 @@ import (
 	"errors"
 	"fmt"
 	"reflect"
+	"unicode"
 
 	"github.com/bep/gitmap"
 
@@ -488,17 +489,23 @@ var (
 // whether the contentis truncated or not.
 // Note: The content slice will be modified if needed.
 func replaceDivider(content, from, to []byte) ([]byte, bool) {
-	sections := bytes.Split(content, from)
+	dividerIdx := bytes.Index(content, from)
+	if dividerIdx == -1 {
+		return content, false
+	}
+
+	afterSummary := content[dividerIdx+len(from):]
 
 	// If the raw content has nothing but whitespace after the summary
 	// marker then the page shouldn't be marked as truncated.  This check
 	// is simplest against the raw content because different markup engines
 	// (rst and asciidoc in particular) add div and p elements after the
 	// summary marker.
-	truncated := (len(sections) == 2 &&
-		len(bytes.Trim(sections[1], " \n\r")) > 0)
+	truncated := bytes.IndexFunc(afterSummary, func(r rune) bool { return !unicode.IsSpace(r) }) != -1
+
+	content = append(content[:dividerIdx], append(to, afterSummary...)...)
 
-	return bytes.Join(sections, to), truncated
+	return content, truncated
 
 }
 
diff --git a/hugolib/page_test.go b/hugolib/page_test.go
index cc2ab9662ca..7723c02c443 100644
--- a/hugolib/page_test.go
+++ b/hugolib/page_test.go
@@ -1107,7 +1107,7 @@ func TestReplaceDivider(t *testing.T) {
 		expectedTruncated bool
 	}{
 		{"none", "a", "b", "none", false},
-		{"summary divider content", "divider", "HUGO", "summary HUGO content", true},
+		{"summary <!--more--> content", "<!--more-->", "HUGO", "summary HUGO content", true},
 		{"summary\n\ndivider", "divider", "HUGO", "summary\n\nHUGO", false},
 		{"summary\n\ndivider\n\r", "divider", "HUGO", "summary\n\nHUGO\n\r", false},
 	}