First of all I would like to congratulate the authors on their full description of the ISG which will certainly help to increase the visibility of this important IAG service. Hopefully, the DOI assignments will increase the numbers of submitted and publicly available geoid models. Overall, the description of the ISG is easy to follow and written in great detail. In my opinion, some details could have been reduced but certainly they will be useful for non-expert users. The authors have carefully revised their manuscript to the suggestions of the two referees, whom I would like to thank you for their work.

After considering some minor issues, the manuscript can be published in ESSD. Please find my comments as follows:
• Explanations of abbreviations e.g. ISG and IAG appear more than once. Please check throughout the manuscript.
• l. 49: “not included into global models”
• l. 52: Start a new paragraph with „It is worth…“
• l. 54: “Bouguer”
• l. 54: “can be converted into geoid undulations” (not “into the geoid”)
• l. 136: “not only depend on the amount and quality…”
• l. 139: The structure is not well chosen. Subsection 3.1.1 does not fit into the scheme as there is no 3.1.2 etc. I suggest removing the subsection title and add the content to section 3.1.
• l. 191 and l. 196: Do you really need separate subsections for Data Citation Index and DOI service? I suggest removing the subsection titles and the contents to section 3.2.
• l. 272: Table 1: The last geoid school was held in 2016. This is the largest break of 5 years. Are there any plans regarding a future school?
• l. 309: “long wavelength gravity field variations”
• ll. 318-: What is the added value of section 4.2? Both reviewers stated that chapter 4 is very long. I am also asking for removal of this section. I would follow reviewer 2 stating that one comparison between local geoid models and one between a local and a global model is sufficient. In addition I suggest reducing the title. “by different authors" and “with different techniques” do not change the methodology as aimed here.
• l. 337: “geoid heights”  “geoid undulations” (always used elsewhere)
• l. 411: “are quite in agreement  “agree quite well”
• l. 412: “differences of a few centimeters at most”
• l. 412: “increases by one order of magnitude”
• l. 414: “of the geoid residuals”
• l. 430: “with residuals of about 2.5 cm.” What does this mean? 1 sigma? Min/Max? What are the statistics between GNSS/levelling and the global models?
• l. 431: “suffer of missing information in this area”  “limitation of the terrestrial datasets for this area included in the global models”
• l. 431: Maybe add 1 sentence on the small differences between global models as these contain the same terrestrial data sets.
• l. 450: Start a new (and the last) paragraph regarding future plans.