How can you tell whether SEO is making progress?

Review search performance, real enquiries and completed work together, rather than one ranking number.

SEO (search engine optimisation) progress cannot be judged only by one term’s position. You also need to know whether visitors are suitable, whether they make real enquiries, and what changed on the website.

Start with these basics

Impressions, clicks and enquiries are different stages. In everyday marketing discussions, a “conversion” usually means a visitor completed a desired action, such as submitting an enquiry. It does not necessarily mean a sale.
In Google Analytics, Google’s website analytics tool, important actions are usually recorded as “key events”: actions you want to track closely. These differ from “conversions” in advertising reports. See Google: Conversions and key events .
Branded queries include your business name. Non-branded queries search for a service or question without it. Separating them helps distinguish people who already know you from other discovery opportunities, but does not identify people precisely.
Compare matching periods and scope, allowing for holidays, seasonal demand and website changes. Investigate data and page issues after a sudden decline before assuming a penalty.

See Google: Search performance report and Google: Debug search traffic drops .

Keep a timeline before judging causes

Record dates for page launches, address changes, content updates, advertising, social promotion and holidays. Compare them with impressions, clicks and enquiries. If a page changed and a promotion ran in the same week, increased visits alone cannot prove that either improved organic rankings: positions in unpaid search results. In someone else’s case study, a domain’s registration date may not be the website’s launch date. Historical page snapshots show only what could be seen at that time.

Third-party estimates of competitors’ traffic and rankings are usually estimates, rather than actual figures from their accounts. If tools suddenly disagree, check the period, region, domain version and tool documentation, then return to specific searches and pages. See the investigation approach in Google: Debug search traffic drops .

If a new domain grows suddenly, check whether an older website moved to it or an existing brand, parent website or customer base brought visits. Review old and new addresses, redirects that send visitors to the new pages, and verifiable publication records. A new URL does not mean starting from zero. Even a confirmed migration may not explain all growth; investigate the remaining changes too. See Google: Site moves .

Keep visits, return visits and enquiries separate

Even when organic search accounts for a large share of visits, check actual clicks and queries. Branded searches may follow earlier advertising, recommendations or existing customer relationships. When studying non-branded opportunities, check the business name, common spacing variations and misspellings, then see which service or question pages people discover. A growing percentage can also mean other channels have declined.

“Direct” traffic usually means the tool could not identify a clear source. It can include typed addresses and bookmarks, but also shared links with missing source information. It cannot all be counted as loyal return visits or brand awareness. See Google: Direct traffic .

PV means page views: how often pages were viewed. UV commonly means the distinct visitors a tool identifies; identification methods vary. A visit can include multiple pages. Adding detail pages or navigation can change pages per visit, so record the redesign date. More visits and longer stays do not necessarily mean satisfied customers; people may be struggling to find an action. Compare enquiries, repeated use and real feedback before judging the experience.

Revenue cannot be calculated directly from visit counts. Record enquiries, suitable customer needs, sales rates and actual charges. Do not apply another case study’s assumed conversion rate, visit frequency or price directly to your business.

A high event ranking, many participants or many likes may still bring unsuitable visitors. Record arrivals, use, enquiries and sales separately, and ask why participants came. Someone answering a quiz for a gift may not need a repair service. An order after a cheap click may also come from someone who already knew you. The final click’s cost is not the full cost of acquiring that customer.

For advertising, email or partner promotions, you can use UTM parameters: source, channel and campaign labels added to a URL. Test that they record correctly. They help separate labelled sources, but cannot reconstruct every customer’s complete history of contact with you. Unlabelled visits cannot automatically be counted as SEO. See Google: Collect campaign data with URL tags .

Name searches can also include ordinary words, unrelated things with the same name or misspellings that existed before your product launched. Check earlier trends and the meaning of actual queries before counting all those visits as awareness created by your promotion.

Ask whether records are complete before interpreting them. Exports may have row limits or lack older records. Missing exported activity does not mean it never happened. Small samples can mislead too: suppose one of 39 registered users pays. That is one observation, rather than a promise of a stable future rate. One person buying repeatedly is also not several paying customers.

Normal time spent varies by task. Someone who quickly checks opening hours and leaves may have solved their problem. A practice tool may be used for longer. Check task completion before interpreting time spent and return visits.

Ask who was counted in a case study

“Successful pages usually rank within a few months” may describe only a small set of pages that succeeded, excluding the others. Check the sample and the pages without results before applying the finding. It is not a ranking countdown for your website.

Observation (illustrative)	Initial misunderstanding	Interpretation after checking
5 of 100 pages have meaningful clicks	Use those 5 pages’ average time as the expectation for every page	Include the other 95 pages in the assessment
Half of visits are direct	Half are returning customers	Investigate missing sources, then actual repeated use
Pages per visit rise after adding detail pages	The content must be better	Compare structure, task results and customer feedback

Suppose a new website publishes 100 pages and only 5 receive meaningful clicks. The owner should compare demand and content for those 5 pages with the others before planning the next round. Successful pages’ speed cannot promise that every page will succeed.

Walk through an example

Check	Before (illustrative)	After, or the appropriate action
Relevant search clicks	50	65; an increase of 15
Enquiries confirmed as coming from search	4	5; an increase of 1
Work record	Service coverage unclear	Added actual coverage and booking information

Suppose search clicks rise from 50 to 65 after a change, and enquiries confirmed as coming from search rise from 4 to 5. Record those changes. One before-and-after comparison cannot prove that the change caused all of the increase.

Common misunderstandings

Enquiry sources need actual records. Do not credit every enquiry to SEO. More visits can also come from unrelated queries, so a larger number may not provide more business value.