Skip to content

What real website search growth looks like

Two Bangkok website projects, the monthly search figures, and the questions to ask before believing a dramatic marketing result.

buyer guideupdated 15 October 20266 minute readtopics: Academy, SEO, measurement, Search Console, case studies, Thailand

The short answer

UnReal Gym recorded 94% more Google impressions in September than June 2026. Pro Pick Property recorded 3.4x as many in September as July. Those are complete calendar-month comparisons, not promises of sales. Here are the underlying counts, the September declines from August, and what these two selected DLVX cases do and do not establish.

Google Search Console web search, all countries and devices. Complete calendar months, source exported 14 October 2026.
Website and monthDaysImpressionsClicks
UnReal, June 20263016,5621,060
UnReal, July 20263123,5011,284
UnReal, August 20263137,2741,712
UnReal, September 20263032,1551,582
Pro Pick, July 2026311,61545
Pro Pick, August 2026315,900106
Pro Pick, September 2026305,52699

A useful result has a starting point.

A large percentage is not enough to judge a marketing result. You need the original count, the ending count, the dates and an explanation of what was measured. You also need to know what the result does not prove. The table makes those checks possible without asking you to trust a headline.

These are two selected websites DLVX works on in Bangkok. They are first-party case evidence, not an independent audit, a survey of Thai businesses or a ranking of agencies. DLVX holds equity in UnReal, as disclosed in its case study. Neither project is an independent endorsement of this analysis.

Pro Pick: 3.4x the monthly search appearances.

Pro Pick Property's impressions increased from 1,615 in July to 5,526 in September, or 3.4x. That is a 242% increase. Clicks increased from 45 to 99, a 120% increase. July is the first complete calendar month in the available source export, so it is the baseline rather than the eight recorded days in June.

Comparing a short launch window with a full month would inflate the headline. July has 31 days and September has 30, so these remain calendar-month totals, not identical-duration or seasonally adjusted periods. For operating decisions, also check equal-length windows and daily rates.

September was below August for both impressions and clicks. That belongs in the story. A total number of records in a CRM is another measure entirely: it cannot be relabelled as leads generated by the website, SEO or AI without attribution evidence.

UnReal: 94% more impressions against June.

UnReal Gym's impressions increased from 16,562 in June to 32,155 in September, a 94% increase. Clicks increased from 1,060 to 1,582, a 49% increase. Both comparison months contain 30 days. The three recorded May days are excluded from the baseline.

September was below August here too. The website and connected customer journey are explained in the UnReal case, linked below. Being found in search and completing a registration or payment are different steps. A useful measurement plan follows both instead of making one stand in for the other.

Impressions are not enquiries.

Search Console records search appearances and clicks from Google results. An impression is not a unique person or a website visit. A click does not tell you whether the visitor became a customer. Google's metric definitions are linked below so the meaning is not left to the agency presenting the number.

Keep the chain separate: search appearance, click, relevant visit, enquiry, qualified opportunity and sale. Join those stages where tracking permits it and mark the gaps where it does not. These figures show observed search movement. They do not isolate DLVX's contribution, establish incremental revenue or demonstrate that AI caused the growth.

Ask these questions before buying the result.

A useful agency can explain its role, show the measurement window and tell you where the evidence stops. When comparing Bangkok web agencies, use the same questions for every supplier rather than rewarding whichever presentation has the largest multiplier.

  • Which website property and search type were measured?
  • What were the original counts and exact comparison dates?
  • Does either period contain only part of a month?
  • Are country, device and other filters consistent?
  • What happened in the months between the endpoints?
  • Which enquiries or sales were actually attributed, and which remain unknown?

How to reproduce the arithmetic.

The downloadable JSON contains the complete monthly counts in the table, source retrieval timestamp, scope, definitions and limitations. Its source is a daily Google Search Console export retrieved on 14 October 2026 using final web-search data, all countries and all devices, grouped by date at property level. Daily counts were added into calendar months. Data coverage does not establish a project's launch date.

Divide ending impressions by baseline impressions to obtain the multiplier. Subtract one and multiply by 100 for percentage growth. Click growth uses the same formula. Published percentages are rounded. Partial first months are excluded, and the private daily export is not included in the public download.

If your website needs both better discovery and a clearer customer journey, the Bangkok web design and development service linked below explains the next step. Start with what the website needs to do for the business, then agree how to measure whether it does it.

Sources and downloads.

Sources are linked directly. Market prices and engine answers can change, so use the checked date when citing this page.

Need the answer applied to your business?

Send the current site, the decision and any proposal you are comparing. We will tell you what we would keep, what we would challenge and whether DLVX is a fit.

Start with the evidence