You have a CSV file of keywords with two thousand rows, and the monthly search volume for each one is written next to it. After three weeks you produce content and see in Search Console that the very keywords with high volume have received zero clicks. The problem is neither your content nor your site speed; the problem is that those two thousand rows weren't built for the Persian-speaking market. Keyword research in Persian has an extra step that doesn't exist in English guides: before anything else, you must understand where that search volume came from and who is typing it.
Why foreign tools show Persian search volume incorrectly
International tools usually estimate search volume from ad click data and limited samples. For languages like English that have large samples, the error is tolerable. For Persian, the sample is small and the estimation model rounds the number. The result is that a keyword with a real volume of 40 searches per month is shown as 720 in the tool, and another keyword with 900 searches is shown as 210.
The solution is to replace the data source. Search Console gives real impression and click data for any site that has been registered there for at least a few months. For new keywords, use Google autocomplete suggestions and the "related searches" section; these are real behavioral data from Iranian users, not estimates. If you want to look at several sources side by side, SEO and webmaster tools is a good starting point, but none of them replaces your own Search Console data.
A quick method for gauging real volume
Search the keyword on Google and look at the number of results and the date of the latest content on the first page. If the first ten results all belong to past years and no new sites are ranking there, the search volume is probably low or the search intent is saturated. Conversely, if three new sites with weak content are ranking on the first page, it means real competition is low and the value of producing content is high. This test takes thirty seconds and is more accurate than reading a search volume table.
The high-volume, zero-intent trap: keywords that get no clicks
A large portion of Persian searches are "finding a specific thing" searches: the name of a movie, the price of a phone, the working hours of an office, the tracking number of a shipment. These keywords have high volume and look golden in tools. But a user searching "Samsung phone price" isn't looking for your article; they're looking for a number. If you write a 1,500-word piece, even if you rank, the bounce rate will be above 90 percent, and that same signal tells Google this page didn't give the right answer.
You see the sign in Search Console: a page with high impressions, CTR below one percent, and an average position between 5 and 10. This combination almost always means you've misread the search intent, not that the content is weak.
Before producing content, place each keyword in one of these four buckets:
- Informational: "how to change a domain's DNS" — an article, guide, or video answers it.
- Comparative: "difference between NVMe and SSD" — it needs a table and analysis.
- Transactional: "buy WordPress hosting" — a product page, not a blog.
- Navigational: "ServerNet panel login" — the user is looking for a specific address; producing content for this bucket is almost always a waste.
Delete the fourth bucket. Send the third bucket to a sales page. Write articles only for the first two buckets.
Here's where they go wrong: lumping same-root keywords onto one page
A pattern I've seen a lot: a page is built with the title "hosting, buy hosting, cheap hosting, WordPress hosting, Linux hosting" and it's assumed this page covers all those keywords. The result in Search Console is this: the page isn't on the first page for any of those phrases, but five other pages of the site, each of which addressed one specific topic, have lost rankings. Google can't tell what this page is about, and the other pages lose their topical signal too.
The practical sign is this: in the Search Console "Coverage" report, you see a page that has impressions for dozens of phrases but its average position is above 30. Break that page apart; each important phrase gets its own page with its own title.
The problem of Arabic and English synonyms
In Persian, one concept is written in several forms: "وبسایت" and "وب سایت" and "website", "ایمیل" and "email". If you only include one in your content, you lose part of the search. But stuffing all the forms into one paragraph makes the text unreadable for the reader. The right way: use the main form in the title and H2s, bring the other forms once in the text, and leave the rest to internal links and anchor text.
From keyword list to site structure: practical clustering
After cleanup, you usually go from two thousand keywords to 150 to 250 usable keywords. Cluster these based on the similarity of Google results, not lexical similarity. Simple test: search two keywords separately; if their first ten results share more than six items, one page is enough for them. If not, you need two pages.
Each cluster gets a pillar page and longer keywords become its sub-topics. For example, the "site speed" cluster becomes the pillar and "high TTFB", "removing render blocking", and "image optimization" become sub-topics. This structure is readable both for Google and for you: when the pillar page ranks, the sub-pages rise faster. You can see an example in the explanation of TTFB and how it differs from load time, which is itself a sub-topic of that same cluster.
Long-tail keywords that are actually worth it
A three-word keyword with a volume of 30 searches per month, if it has purchase intent, is more valuable than a single-word keyword with a volume of 5,000 and zero intent. In the Persian market this ratio is more extreme, because competition on short keywords is usually divided among a few large sites and entering it is almost impossible for a new site. Long-tail keywords, on the other hand, often have a first page with only two or three relevant results.
Tools: what to measure with what
| Source | What it gives | Where it misleads |
|---|---|---|
| Search Console | Real impressions and clicks, average position | Only keywords the site has already had impressions for |
| Google autocomplete | Real user phrases | No volume number; must be prioritized manually |
| Foreign search volume tools | Quick estimate for relative comparison | Absolute numbers aren't reliable for Persian |
| Site internal search report | The real intent of your own visitors | Only visitors who have already found the site |
Don't underestimate the last row. If your site has internal search, open its log. The keywords users type inside your site and don't get an answer to may be the best possible keyword list, because both the intent is clear and you've measured the competition.
The cost no one accounts for
Precise keyword research is slow. For 200 keywords, manually testing Google results and reading the intent of each takes at least two working days. In contrast, grabbing a CSV file and starting content production takes half an hour. If you have a small team and publish four articles a month, those two days are a logical investment. But if you publish three articles a day, this method becomes a bottleneck and you need to move to a lighter process: manually review only the 30 main keywords and guess the rest based on clusters.
There's another limitation that's less often mentioned: Search Console data for keywords with low impressions isn't shown due to the privacy threshold. That means precisely the long-tail, low-volume keywords that are valuable for a new site aren't visible in Search Console. For these, you must rely on Google autocomplete and manual analysis of the results page.
Before writing the first word, check these
If the site has just launched, first make sure a technical problem isn't blocking indexing; otherwise even the best keyword research is useless. Start with technical domain and DNS check and look at chained redirects and robots settings. Then move on to content.
And if your research output leads to a site structure redesign, do this before mass content production. Moving 50 articles after publication comes with redirects and a temporary ranking drop. The ServerNet SEO services team usually does this step first, because fixing the structure on an empty site is cheaper than on a full one.
The next step is clear: open the keyword file and temporarily delete the search volume column. Instead, add a column titled "intent". Any row you can't place in one of the three useful buckets within ten seconds should be deleted right there. What remains is your real work list.
Frequently asked questions
Can I do keyword research only with Google Trends?
No, Google Trends doesn't give absolute search volume and only shows relative trends. It's good for seeing the seasonal popularity of a phrase, but for deciding on content production it must be combined with Search Console data or manual testing of the results page.
Why does a keyword with high search volume get no clicks on my site?
Usually the search intent doesn't match the page type. If the user is looking for a price or a phone number and you've provided an article, you won't get clicks even with a good ranking. In Search Console, look at the combination of high impressions and very low CTR; this is the main sign.
For a Persian site, what monthly search volume is worth producing content for?
There's no fixed number, but for a new site, a keyword with 50 to 200 monthly searches and clear informational intent usually yields a better return than a 5,000-search keyword with heavy competition. Make intent and the difficulty of the results page your criterion, not volume.
How many keywords should I fit into one article?
One main keyword for the title and main H2s, and two to four same-cluster keywords that fit naturally in the text. More than this turns the text into a keyword list and creates the same problem you saw in the same-root keyword section.
Comments 0
No comments yet — be the first!