diff --git a/README.md b/README.md
index 13f6986b..6dc4200e 100644
--- a/README.md
+++ b/README.md
@@ -4,7 +4,7 @@
 
 CUTLASS is a collection of CUDA C++ template abstractions for implementing 
 high-performance matrix-multiplication (GEMM) at all levels and scales within CUDA. 
-It incorporates the same strategies for data movement and hierarchical decomposition 
+It incorporates the same strategies for hierarchical decomposition and data movement 
 that are used to implement cuBLAS.  CUTLASS decomposes these “moving parts” into 
 reusable, modular software components abstracted by C++ template classes.  These
 thread-wide, warp-wide, block-wide, and device-wide primitives can be specialized