diff --git a/benchmark.py b/benchmark.py
--- a/benchmark.py
+++ b/benchmark.py
@@ -18,7 +18,9 @@
 SEED_OPTIONS = ["-s 10"]
 
 
-def get_scenario_cmd(algo, kb_url, kb_label, origin_info, extracted_repo_path):
+def get_scenario_cmd(
+    algo, kb_url, kb_label, origin_info, extracted_repo_path, log_file
+):
     return [
         "swh",
         "scanner",
@@ -33,6 +35,8 @@
         origin_info["origin"],
         "--commit",
         origin_info["commit"],
+        "--log-file",
+        log_file,
         "--exclude",
         str(extracted_repo_path) + "/.git",
         str(extracted_repo_path),
@@ -40,7 +44,7 @@
 
 
 def run_experiments(
-    repo_path: str, temp_path: str, kb_state_file: str, algos: Set[str]
+    repo_path: str, temp_path: str, kb_state_file: str, algos: Set[str], log_file: str
 ):
     """This function create a process for each experiment; one experiment is composed
     by: the repository we want to scan, the algorithms we need to test and different
@@ -70,13 +74,23 @@
             if algo == "random":
                 for seed_opt in SEED_OPTIONS:
                     random_cmd = get_scenario_cmd(
-                        algo, kb_url, kb_label, origin_info, str(extracted_repo_path)
+                        algo,
+                        kb_url,
+                        kb_label,
+                        origin_info,
+                        str(extracted_repo_path),
+                        log_file,
                     )
                     scenario_cmds.append(random_cmd + [seed_opt])
             else:
                 scenario_cmds.append(
                     get_scenario_cmd(
-                        algo, kb_url, kb_label, origin_info, str(extracted_repo_path)
+                        algo,
+                        kb_url,
+                        kb_label,
+                        origin_info,
+                        str(extracted_repo_path),
+                        log_file,
                     )
                 )
 
@@ -97,6 +111,9 @@
 @click.argument("repo_path", type=click.Path(exists=True), required=True)
 @click.argument("temp_path", type=click.Path(exists=True), required=True)
 @click.argument("kb_state", type=click.Path(exists=True), required=True)
+@click.option(
+    "--log-file", "-l", metavar="FILENAME", required=True, help="Log custom path."
+)
 @click.option(
     "-a",
     "--algo",
@@ -110,9 +127,9 @@
     metavar="ALGORITHM_NAME",
     help="The algorithm name for the benchmark.",
 )
-def main(repo_path, temp_path, kb_state, algos):
+def main(repo_path, temp_path, kb_state, log_file, algos):
     logging.basicConfig(
-        filename="experiments.log",
+        filename=log_file,
         format="%(asctime)s %(message)s",
         datefmt="%m/%d/%Y %I:%M:%S %p",
     )
@@ -125,7 +142,7 @@
                 stdout=subprocess.DEVNULL,
                 stderr=sys.stderr,
             )
-            run_experiments(repo_path, temp_path, kb_state, set(algos))
+            run_experiments(repo_path, temp_path, kb_state, set(algos), log_file)
     except Exception as e:
         logging.exception(e)
     except IOError as ioerror:
diff --git a/run_benchmark.sh b/run_benchmark.sh
--- a/run_benchmark.sh
+++ b/run_benchmark.sh
@@ -13,6 +13,7 @@
 
 temp_dir=$1
 kb_state=$2
+log_file=$3
 
 if [ ! -d "$temp_dir" ]; then
     echo "You should provide a valid temporary directory path"
@@ -24,7 +25,14 @@
     exit 1
 fi
 
-for i in "${@:3}"; do
+if [ "$log_file" == '' ]; then
+    echo "You should provide the file where the logs will be saved"
+    exit 1
+else
+    log_file="-l ${log_file}"
+fi
+
+for i in "${@:4}"; do
     algos="$algos -a $i"
 done
 
@@ -33,5 +41,5 @@
 
 while IFS= read -r repo;
 do
-    ./benchmark.py $repo $temp_dir $kb_state $algos
+    ./benchmark.py $repo $temp_dir $kb_state $algos $log_file
 done
diff --git a/swh/scanner/benchmark_algos.py b/swh/scanner/benchmark_algos.py
--- a/swh/scanner/benchmark_algos.py
+++ b/swh/scanner/benchmark_algos.py
@@ -6,6 +6,7 @@
 import collections
 import itertools
 import json
+import logging
 import os
 from pathlib import Path
 import random
@@ -353,6 +354,10 @@
     counter["api_calls"] = 0
     counter["queries"] = 0
     source_tree = load_source(Path(root), sre_patterns)
+    logging.info(
+        f'started processing repo "{repo_id}" with algorithm '
+        f'"{algo}" and knowledge base "{backend_name}"'
+    )
 
     if algo == "random":
         if seed:
@@ -393,4 +398,9 @@
         counter["queries"],
     )
 
+    logging.info(
+        f'finished processing repo "{repo_id}" with algorithm '
+        f'"{algo}" and knowledge base "{backend_name}"'
+    )
+
     print(*result, sep=",")
diff --git a/swh/scanner/cli.py b/swh/scanner/cli.py
--- a/swh/scanner/cli.py
+++ b/swh/scanner/cli.py
@@ -281,9 +281,19 @@
 @click.option(
     "--seed", "-s", metavar="SEED", type=int, help="Seed for the random algorithm"
 )
+@click.option("--log-file", "-l", metavar="FILENAME", help="Log custom path.")
 @click.pass_context
 def benchmark(
-    ctx, root_path, api_url, backend_name, origin_url, commit, patterns, algo, seed
+    ctx,
+    root_path,
+    api_url,
+    backend_name,
+    origin_url,
+    commit,
+    patterns,
+    algo,
+    seed,
+    log_file,
 ):
     from importlib import reload
     import logging
@@ -293,9 +303,10 @@
     # reload logging module avoid conflict with benchmark.py logging
     reload(logging)
     logging.basicConfig(
-        filename="experiments.log",
+        filename=log_file if log_file else "experiments.log",
         format="%(asctime)s %(message)s",
         datefmt="%m/%d/%Y %I:%M:%S %p",
+        level=logging.INFO,
     )
 
     try: