diff --git a/phlex/core/edge_creation_policy.hpp b/phlex/core/edge_creation_policy.hpp
index 98e5e426..c5d2d85d 100644
--- a/phlex/core/edge_creation_policy.hpp
+++ b/phlex/core/edge_creation_policy.hpp
@@ -45,9 +45,6 @@ namespace phlex::experimental {
     std::multimap<product_suffix_t, named_output_port> result;
     for (auto const& [node_name, node] : nodes) {
       for (auto const& product_spec : node->output()) {
-        if (product_spec.suffix().empty())
-          continue;
-
         result.emplace(product_spec.suffix(),
                        named_output_port{node_name, &node->output_port(), product_spec.type()});
       }
diff --git a/phlex/model/product_specification.cpp b/phlex/model/product_specification.cpp
index 6f018ce7..9768beaf 100644
--- a/phlex/model/product_specification.cpp
+++ b/phlex/model/product_specification.cpp
@@ -2,6 +2,8 @@
 #include "phlex/model/algorithm_name.hpp"
 
 #include <cassert>
+#include <ranges>
+#include <stdexcept>
 #include <utility>
 
 namespace phlex::experimental {
@@ -48,19 +50,29 @@ namespace phlex::experimental {
     return {algorithm_name::create(""), identifier(s), type_id{}};
   }
 
-  product_specifications to_product_specifications(std::string_view const name,
+  product_specifications to_product_specifications(std::string_view const algorithm_specification,
                                                    std::vector<std::string> output_suffixes,
                                                    std::vector<type_id> output_types)
   {
-    assert(output_suffixes.size() == output_types.size());
-    product_specifications outputs;
-    outputs.reserve(output_suffixes.size());
+    static std::string const default_product_suffix = "";
 
-    // zip view isn't available until C++23 so we have to use a loop over the index
-    for (std::size_t i = 0; i < output_suffixes.size(); ++i) {
-      outputs.emplace_back(
-        algorithm_name::create(name), identifier(output_suffixes[i]), output_types[i]);
+    if (output_suffixes.empty()) {
+      // This happens whenever the user does not specify output_product_suffixes(...), so we
+      // assign the default suffix to all output data products.
+      output_suffixes.assign(output_types.size(), default_product_suffix);
     }
-    return outputs;
+
+    // If output_suffixes and output_types are non-empty, the framework guarantees that they will
+    // have the same length. They will either have the same length due to the default-suffix
+    // assignment above, or because the call to output_product_suffixes(...) demands it.
+    assert(output_suffixes.size() == output_types.size());
+
+    // We can use std::views::zip_transform once the AppleClang C++ STL supports it.
+    auto const algo_name = algorithm_name::create(algorithm_specification);
+    return std::views::zip(output_suffixes, output_types) |
+           std::views::transform([&algo_name](auto const& p) {
+             return product_specification{algo_name, identifier(std::get<0>(p)), std::get<1>(p)};
+           }) |
+           std::ranges::to<product_specifications>();
   }
 }
diff --git a/phlex/model/product_specification.hpp b/phlex/model/product_specification.hpp
index 37f95ce9..1eeb93c8 100644
--- a/phlex/model/product_specification.hpp
+++ b/phlex/model/product_specification.hpp
@@ -38,7 +38,7 @@ namespace phlex::experimental {
 
   private:
     algorithm_name qualifier_;
-    identifier suffix_;
+    identifier suffix_; // Default suffix is empty string
     type_id type_id_{};
   };
 
diff --git a/test/hierarchical_nodes.cpp b/test/hierarchical_nodes.cpp
index 70f87147..025970aa 100644
--- a/test/hierarchical_nodes.cpp
+++ b/test/hierarchical_nodes.cpp
@@ -101,23 +101,19 @@ TEST_CASE("Hierarchical nodes", "[graph]")
 
   g.transform("get_the_time", strtime, concurrency::unlimited)
     .input_family(product_query{.creator = "input", .layer = "run", .suffix = "time"})
-    .experimental_when()
-    .output_product_suffixes("strtime");
+    .experimental_when();
   g.transform("square", square, concurrency::unlimited)
-    .input_family(product_query{.creator = "input", .layer = "event", .suffix = "number"})
-    .output_product_suffixes("squared_number");
+    .input_family(product_query{.creator = "input", .layer = "event", .suffix = "number"});
 
   g.fold("add", add, concurrency::unlimited, "run", 15u)
-    .input_family(product_query{.creator = "square", .layer = "event", .suffix = "squared_number"})
-    .experimental_when()
-    .output_product_suffixes("added_data");
+    .input_family(product_query{.creator = "square", .layer = "event"})
+    .experimental_when();
 
   g.transform("scale", scale, concurrency::unlimited)
-    .input_family(product_query{.creator = "add", .layer = "run", .suffix = "added_data"})
-    .output_product_suffixes("result");
+    .input_family(product_query{.creator = "add", .layer = "run"});
   g.observe("print_result", print_result, concurrency::unlimited)
-    .input_family(product_query{.creator = "scale", .layer = "run", .suffix = "result"},
-                  product_query{.creator = "get_the_time", .layer = "run", .suffix = "strtime"});
+    .input_family(product_query{.creator = "scale", .layer = "run"},
+                  product_query{.creator = "get_the_time", .layer = "run"});
 
   g.make<experimental::test::products_for_output>()
     .output("save", &experimental::test::products_for_output::save)
diff --git a/test/provider_test.cpp b/test/provider_test.cpp
index fd02eeb7..1c87b38b 100644
--- a/test/provider_test.cpp
+++ b/test/provider_test.cpp
@@ -43,10 +43,10 @@ TEST_CASE("provider_test")
       product_query{.creator = "input", .layer = "spill", .suffix = "happy_vertices"});
 
   g.transform("passer", pass_on, concurrency::unlimited)
-    .input_family(product_query{.creator = "input", .layer = "spill", .suffix = "happy_vertices"})
-    .output_product_suffixes("vertex_data");
+    .input_family(product_query{.creator = "input", .layer = "spill", .suffix = "happy_vertices"});
 
   g.execute();
+
   CHECK(g.execution_count("passer") == max_events);
   CHECK(g.execution_count("my_name_here") == max_events);
 }