1,500 Mechanical Turk workers were instructed to: search for [money term] search for [money term, test site brand] click on [test site] I wanted to see if a follow-up search and high CTR had any impact on Google Suggest and rankings. While we did notice a spike in clicks and impressions, Google
1,500 Mechanical Turk workers were instructed to:
- search for [money term]
- search for [money term, test site brand]
- click on [test site]
I wanted to see if a follow-up search and high CTR had any impact on Google Suggest and rankings.
While we did notice a spike in clicks and impressions, Google Suggest wasn't influenced in any way and neither were rankings.
A follow-up experiment could include a and consistent longer test period with more workers but it would cost a fair bit, so I'm parking this idea for now.
Extra Information
Amazon doesn't give much detail about qualified workers:
Here's a sample of mturk
workers in csv format.
I do have Search Console data for the test query and here's a typical day:
Locations
| Countries |
Clicks |
| India |
283 |
| United States |
147 |
| Turkey |
11 |
| Macedonia |
10 |
| Unknown Region |
10 |
| Australia |
7 |
| Nigeria |
5 |
| Sri Lanka |
5 |
| United Kingdom |
4 |
| Venezuela |
4 |
| Croatia |
4 |
| Greece |
4 |
| Italy |
3 |
| France |
2 |
| Belize |
2 |
| Philippines |
2 |
| Canada |
1 |
| Romania |
1 |
| Germany |
1 |
| Thailand |
1 |
| Bangladesh |
1 |
| Bahrain |
1 |
| Colombia |
1 |
| Guatemala |
1 |
| Japan |
1 |
| Mexico |
1 |
| Sweden |
1 |
| Tunisia |
1 |
| Spain |
0 |
| Ireland |
0 |
| Iran |
0 |
| Poland |
0 |
| Slovenia |
0 |
Device Breakdown
| Devices |
Clicks |
| Desktop |
486 |
| Mobile |
20 |
| Tablet |
9 |
