WEBVTT

1
00:00:00.000 --> 00:00:02.480
Keyword research is not finished when the spreadsheet is full.

2
00:00:02.500 --> 00:00:09.000
It is finished when a real search job has a useful page, the website can support that page, and you know what to measure next.

3
00:00:09.400 --> 00:00:13.580
That is the workflow I use for keyword research and website SEO.

4
00:00:13.880 --> 00:00:18.580
Start with one audience, one business outcome, and a bounded seed topic.

5
00:00:18.940 --> 00:00:22.280
Gather keyword evidence, but keep four different layers separate.

6
00:00:22.900 --> 00:00:25.600
Market estimates tell you what demand might exist.

7
00:00:25.960 --> 00:00:28.690
Live search results show what Google is serving now.

8
00:00:28.700 --> 00:00:33.060
Site diagnostics show whether your page is technically eligible and usable.

9
00:00:33.840 --> 00:00:36.400
First-party outcomes show what actually happened on your website.

10
00:00:36.520 --> 00:00:38.040
Those layers help each other.

11
00:00:38.240 --> 00:00:39.920
They do not replace each other.

12
00:00:40.420 --> 00:00:42.120
Search volume is not traffic.

13
00:00:42.500 --> 00:00:44.980
A difficulty score is not a Google metric.

14
00:00:45.160 --> 00:00:47.860
A clean audit score is not a ranking promise.

15
00:00:48.340 --> 00:00:52.020
Next, group keywords by the job a person is trying to complete.

16
00:00:52.360 --> 00:00:54.970
One useful page can serve a compatible cluster.

17
00:00:54.980 --> 00:01:00.680
A new page is only justified when the audience, task, or required format is genuinely different.

18
00:01:00.720 --> 00:01:03.060
Otherwise you are not building topical authority.

19
00:01:03.240 --> 00:01:07.620
You are giving the same idea several URLs and asking them to compete for attention.

20
00:01:08.120 --> 00:01:10.500
Then map each cluster to a decision.

21
00:01:10.900 --> 00:01:17.260
Keep the current page, refresh it, merge overlapping pages, create something new, or decline the opportunity.

22
00:01:17.500 --> 00:01:19.380
Declining a shiny keyword is allowed.

23
00:01:19.400 --> 00:01:20.390
The spreadsheet will recover.

24
00:01:20.400 --> 00:01:23.420
Once the page purpose is clear, audit the destination.

25
00:01:23.540 --> 00:01:30.820
Check discovery, crawlability, indexability, canonical signals, rendering, internal links, content quality, accessibility, and performance.

26
00:01:31.340 --> 00:01:37.980
Prioritize fixes by impact, confidence, effort, and risk, not by whichever score has the most dramatic color.

27
00:01:38.100 --> 00:01:41.290
The Research Pro walkthrough grounds this in one recorded run.

28
00:01:41.300 --> 00:01:46.040
The keyword export expanded ai marketing into 282 rows.

29
00:01:46.080 --> 00:01:53.360
The website report covered 42 crawled pages and showed a composite health score of 98 alongside 13 findings.

30
00:01:53.880 --> 00:01:57.460
Those values belong to that dated provider-backed example.

31
00:01:57.860 --> 00:02:02.320
They are evidence of what the run displayed, not a forecast for another site.

32
00:02:02.460 --> 00:02:09.300
Finally, write down the page, intended search job, exact change, date, and measurement plan before publishing.

33
00:02:09.840 --> 00:02:15.220
Then use Search Console, analytics, and real business outcomes to challenge the decision.

34
00:02:15.680 --> 00:02:18.520
The useful output is not a bigger keyword list.

35
00:02:18.540 --> 00:02:24.880
It is one defensible page decision, connected to a website that can support it, with evidence that can prove you wrong later.

36
00:02:25.400 --> 00:02:28.280
That is less magical than pressing an optimize button.

37
00:02:28.500 --> 00:02:33.660
It is also how SEO becomes a learning system instead of a very confident-looking pile of columns.