What MYNIC's August and September totals can tell domain buyers, and why a falling count cannot establish a renewal rate or buying opportunity.
A falling domain-registration total can look like a renewal verdict. It is not one. Two snapshots tell you how the reported total changed, but not which names renewed, which expired, or why their owners made those decisions. That distinction matters when you use extension statistics to research domains. A decline may deserve a closer look.
It does not, by itself, establish a renewal rate, prove weaker buyer demand, or identify names you can register or backorder. What the MYNIC table actually shows MYNIC's 2026 statistics page , checked on October 5, lists a total of 826,384 domain names for August and 693,122 for September. The difference is 133,262 fewer names, or about 16.13% of the August total.
Read the column labels before repeating the numbers. The table's Total combines the listed categories, including .my , .com.my and others. The separate .my column shows 681,864 for August and 548,280 for September. Do not label the combined total as the count in that individual column. These are observations from the registry's table, not a list of expired names.
The table does not identify the same group of domains approaching renewal or report their individual outcomes. It therefore does not establish that 16.13% of names due for renewal failed to renew. Keep three questions separate Stock: How many names were counted at an observation point, within the source's stated categories? Change: How did that total differ between two observations?
A net change does not separately measure additions and removals. Cohort renewal: What happened to a defined group of names due for renewal, measured under a stated rule and observation window? A renewal percentage needs a meaningful denominator.
If you want to measure first-renewal behavior for names registered during a promotion, you need evidence about that particular group, when its renewal decisions fall due, and which outcomes count as renewed. A later registry-wide total includes a different population. Registration terms, observation dates and the treatment of pending outcomes also matter.
Record those definitions instead of assuming that every name in a starting total reaches a renewal decision during the following month. A hypothetical example shows the missing information Imagine a consistently defined count starts at 1,000 names and ends at 900. One possible path is 200 additions and 300 removals. Another is 50 additions and 150 removals. Both produce the same net decline of 100.
Those are invented examples, not MYNIC data. They show why subtracting two totals cannot recover separate flows. Neither example tells you which names belonged to a promotion, which removals followed non-renewal, or how many eligible names renewed. Even a separately reported deletion count would not automatically answer the cohort question. You would still need the population, event definition and observation window.