diff --git a/R/lay.R b/R/lay.R
index 54f204a..b183458 100644
--- a/R/lay.R
+++ b/R/lay.R
@@ -7,7 +7,7 @@
 #' frame (`.data`) as the vector input of some function(s) `.fn`.
 #'
 #' This makes the creation of new columns based on a rowwise operation both simple (see
-#' **Examples**; below) and efficient (see the vignette [**benchmarks**](../doc/benchmark.html)).
+#' **Examples**; below) and efficient (see the Article [**benchmarks**](https://courtiol.github.io/lay/articles/benchmarks.html)).
 #'
 #' The function should be fully compatible with `{dplyr}`-based workflows and follows a syntax close
 #' to [dplyr::across()].
@@ -49,7 +49,7 @@
 #'   - "tidy", internally uses [purrr::pmap()] and is stricter with respect to class coercion
 #'   across columns.
 #'
-#'   The default has been chosen based on these [**benchmarks**](../doc/benchmark.html).
+#'   The default has been chosen based on these [**benchmarks**](https://courtiol.github.io/lay/articles/benchmarks.html).
 #'
 #' @importFrom vctrs vec_c
 #' @importFrom rlang list2 exec as_function
diff --git a/man/lay.Rd b/man/lay.Rd
index 4356c58..d8ca7ae 100644
--- a/man/lay.Rd
+++ b/man/lay.Rd
@@ -31,7 +31,7 @@ is used to apply the rowwise job:
 across columns.
 }
 
-The default has been chosen based on these \href{../doc/benchmark.html}{\strong{benchmarks}}.}
+The default has been chosen based on these \href{https://courtiol.github.io/lay/articles/benchmarks.html}{\strong{benchmarks}}.}
 }
 \description{
 Create efficiently new column(s) in data frame (including tibble) by applying a function one row
@@ -42,7 +42,7 @@ at a time.
 frame (\code{.data}) as the vector input of some function(s) \code{.fn}.
 
 This makes the creation of new columns based on a rowwise operation both simple (see
-\strong{Examples}; below) and efficient (see the vignette \href{../doc/benchmark.html}{\strong{benchmarks}}).
+\strong{Examples}; below) and efficient (see the Article \href{https://courtiol.github.io/lay/articles/benchmarks.html}{\strong{benchmarks}}).
 
 The function should be fully compatible with \code{{dplyr}}-based workflows and follows a syntax close
 to \code{\link[dplyr:across]{dplyr::across()}}.
diff --git a/vignettes/articles/alternatives.Rmd b/vignettes/articles/alternatives.Rmd
index 2ee089f..be90d77 100644
--- a/vignettes/articles/alternatives.Rmd
+++ b/vignettes/articles/alternatives.Rmd
@@ -1,5 +1,5 @@
 ---
-title: "Alternatives to `lay()`"
+title: "Alternatives for rowwise jobs"
 ---
 
 ```{r, include = FALSE}
@@ -12,13 +12,13 @@ knitr::opts_chunk$set(
 ## Article overview
 
 There are many alternatives to perform rowwise jobs in R.
-In this article, we consider, in turns, these alternatives.
+In this Article, we consider, in turns, these alternatives.
 We will stick to our example about drugs usage shown in [introduction](https://courtiol.github.io/lay/).
 The idea is to compare alternative ways to create a new variable named `everused` which indicates if each respondent has used any of the considered pain relievers for non medical purpose or not.
 
 ## Loading packages
 
-This article requires you to load the following packages:
+This Article requires you to load the following packages:
 
 ```{r load_pkg, message=FALSE}
 library(lay)        ## for lay() and the data
@@ -150,6 +150,6 @@ There are probably other ways. If you think of a nice one, please leave an issue
 
 ## Efficiency
 
-The results of a benchmark comparing alternative implementations for our simple rowwise job are shown in another article (see [benchmark](https://courtiol.github.io/lay/articles/benchmark.html)).
+The results of benchmarks comparing alternative implementations for our simple rowwise job are shown in another Article (see [benchmarks](https://courtiol.github.io/lay/articles/benchmarks.html)).
 As you will see, `lay()` is not just simple and powerful, it is also quite efficient!
 
diff --git a/vignettes/articles/benchmark.Rmd b/vignettes/articles/benchmarks.Rmd
similarity index 92%
rename from vignettes/articles/benchmark.Rmd
rename to vignettes/articles/benchmarks.Rmd
index af00aa3..8c9d48c 100644
--- a/vignettes/articles/benchmark.Rmd
+++ b/vignettes/articles/benchmarks.Rmd
@@ -1,5 +1,5 @@
 ---
-title: "Benchmarking alternative implementations for rowwise jobs"
+title: "Benchmarks"
 ---
 
 ```{r, include = FALSE}
@@ -11,7 +11,7 @@ knitr::opts_chunk$set(
 
 ## Article overview
 
-The goal of this article is to compare the performances of `lay()` to alternatives described [here](https://courtiol.github.io/lay/articles/alternatives.html).
+The goal of this Article is to compare the performances of `lay()` to alternatives described [here](https://courtiol.github.io/lay/articles/alternatives.html).
 As you will see, the code using `lay()` is quite efficient.
 The only alternative that is clearly more efficient is the one labeled below "*vectorized*".
 Unfortunately, such a vectorized approach imply to refer explicitly to all column names which data are used.
@@ -21,7 +21,7 @@ and arithmetic operators and does allow the use of other types of functions.
 
 ## Loading packages
 
-This article requires you to load the following packages:
+This Article requires you to load the following packages:
 
 ```{r load_pkg, message=FALSE}
 library(lay)        ## for lay() and the data
@@ -30,8 +30,8 @@ library(tidyr)      ## for pivot_longer() and pivot_wider()
 library(purrr)      ## for pmap_lgl()
 library(slider)     ## for slide()
 library(data.table) ## for an alternative to base and dplyr
-library(bench)      ## for running the benchmark
-library(forcats)    ## for sorting levels in benchmark plot with fct_reorder()
+library(bench)      ## for running the benchmarks
+library(forcats)    ## for sorting levels in plot with fct_reorder()
 ```
 
 Please install them if they are not present on your system.
@@ -55,9 +55,9 @@ drugs_full |>
   pivot_wider(names_from = used, values_from = n)
 ```
 
-In this article, we compare the efficiency of alternative ways to create a new variable named `everused` which indicates if each respondent has used any of the considered pain relievers for non medical purpose or not.
+In this Article, we compare the efficiency of alternative ways to create a new variable named `everused` which indicates if each respondent has used any of the considered pain relievers for non medical purpose or not.
 
-## Benchmark
+## Benchmarks
 
 Let's compare the running time of different methods to do this job:
 
@@ -189,7 +189,7 @@ benchmark2 |>
 
 Note again that the x-axis of the plot is on a logarithmic scale.
 
-## Benchmark environment
+## Benchmarks' environment
 
 ```{r session}
 sessionInfo()