Privacy notice: Apptycoons CMO Publisher
This notice covers the "Apptycoons CMO Publisher", the publishing tool of our internal customer-relationship system at crm.apptycoons.com, and the data it handles. The tool is used only by Apptycoons staff, who sign in with their apptycoons.com Google Workspace accounts. It is not offered to the public. This notice does not cover the contact and lead records our staff keep in the CRM. Our website, and the personal data of people who contact us or work with us, are covered by the apptycoons.com privacy policy.
What the CMO Publisher does
It publishes our own marketing posts and short videos to Apptycoons' own company accounts:
- our LinkedIn company page;
- our YouTube channel (@apptycoons);
- our Facebook Page and Instagram account;
- our X account (@ApptycoonsHQ).
Our staff review and approve each post before it goes out. The tool then reads the statistics of those posts so we can see what works. Some are public counts, such as views, likes, comments and shares. Others are analytics the platforms show only to the account's admins, such as impressions, unique impressions or reach, clicks, saves and watch time. It also reads public posts on topics we work on to suggest post ideas (see "Public posts we read for research").
LinkedIn Community Management API
On LinkedIn the tool is the app "Apptycoons Publisher". An admin of the Apptycoons company page connected it through LinkedIn's Community Management API, and it acts only for that page. It:
- publishes posts to the page: text, a link with its preview, images, or a video with its thumbnail;
- adds a first comment, as the page, with the post's link; if that comment can't be posted, it adds the link to the end of our post's text instead;
- reads the page's recent posts, to match posts we scheduled by hand on LinkedIn to their record in our CRM;
- reads the statistics of each of our posts: impressions, unique impressions, clicks, likes, comments, shares and engagement rate. These are totals for the post, not data about who saw it or engaged with it;
- at sign-in, checks which page the signed-in admin manages.
It does not read member profiles, other pages' posts or anyone's feed, and it stores no member data it gets from LinkedIn: no names, no profile details and no lists of who reacted to or commented on our posts. If it can't tell whether its first comment went through, it reads that post's comments once to look for ours and keeps none of them. It never changes the page's settings or admins.
YouTube API Services
The CMO Publisher uses YouTube API Services to upload our own videos to our own YouTube channel, set their thumbnails, and read their statistics. By using it you agree to the YouTube Terms of Service. Google's handling of data is described in the Google Privacy Policy.
It uploads to and manages only the Apptycoons channel that our staff authorized; it does not manage any other channel. It also uses YouTube API Services to read public data about other public videos for research, as described below. It does not keep data that identifies the viewers of our videos or the people who comment on YouTube; commenter names and IDs in API responses are discarded.
Public posts we read for research
To find topics our buyers care about, the tool reads public posts through official APIs. It does not use LinkedIn's API for this, and its web checks never open LinkedIn pages.
- YouTube, once a week: up to six searches for recent public videos on topics we work on. For the results it reads their titles, channel names, and view and comment counts, each channel's public view and video totals, and the text and like counts of public comments under up to six of those videos. Comments with links are dropped; email addresses, phone numbers and @handles are masked; commenter names and IDs are not kept.
- X, daily and weekly: searches of recent public posts on topics we work on, with each post's text, link, author's @handle and engagement counts, and once a week the follower counts of up to 50 of those authors.
- Public news feeds, Hacker News and GitHub.
The tool sends this research to our AI providers to suggest post ideas. It saves the ideas that pass our checks in the CRM, with links to their sources. An idea built on a public X post keeps that post's text and its author's @handle, and a story idea names its source's author as its credit. A post that answers a public X post may quote it and name its author. The tool also keeps the X user IDs and @handles it has looked up, and for 14 days a list of the sources it turned down.
What we store
In each published post's record in our CRM we store its platform ID, its link, when it went out, when its statistics were last refreshed, and these statistics:
- LinkedIn: impressions, unique impressions, clicks, likes, comments, shares and engagement rate.
- Facebook: views, reach and unique viewers, video plays and replays, average watch time, clicks, reactions by type, comments and shares.
- Instagram: views, reach, likes, comments, saves, shares, total interactions and the post type.
- YouTube: views, likes, comments, favorites and whether the video is public or private.
- X: impressions, likes, reposts, replies, quotes and bookmarks.
From these we also keep a views, reach and engagement total per post, and the platform's error message when a number could not be read. All of them are totals for the post: none says who viewed, clicked, reacted or commented.
We also store:
- The weekly summary of what works that our AI providers write from these numbers. It is saved in the CRM and used when drafting later posts.
- The post ideas the tool saves from its research, described above.
- The access tokens and API keys that let the tool act for our own accounts. They are stored encrypted in AWS Systems Manager Parameter Store. The server holds working copies: setup writes them into its configuration file, which only the server's administrator account can read, and the tool saves the LinkedIn and X tokens it renews in its own data folder on the server. The server's disk and its snapshots are encrypted. Only our server uses these platform tokens, to call the platform that issued them; the one exception is sign-in, when the setup script an administrator runs on their own computer uses a new token once to confirm the account and saves it straight to Parameter Store. The tokens never appear in the CRM.
Cookies
The tool doesn't place or read cookies or any other information on the devices of YouTube viewers, our followers or anyone outside our team. The CRM web app keeps our staff signed in with its own cookies and browser storage; no third party sets cookies or shows ads or content in it.
Who processes this data
We do not sell this data or use it to target ads. These service providers process it for us:
- Amazon Web Services runs our server in Mumbai, India, keeps the tokens in Systems Manager Parameter Store, and stores our encrypted backups in Mumbai and in Frankfurt, Germany.
- OpenAI, and Anthropic when an OpenAI request fails. They draft our posts, and once a week they receive our posts from the last 60 days (each post's channel, format, first line, posting time, statistics, and the number of leads its campaign brought, as a count with no names) to write the summary of what works. They also receive the research described above. If staff ask the CRM's built-in assistant about these records, the records go to the same providers. Under their API terms, this data is not used to train their models.
- GitHub (GitHub Actions) renders our videos from the CRM's post records, using a CRM key kept in GitHub's encrypted secrets.
- ElevenLabs turns our video scripts into voiceovers. It receives only the script text.
- Slack, our internal chat. The tool posts notices about drafts and published posts, with their links, and once a week names our top three posts of the week, with view and engagement counts for Facebook, Instagram and X posts only (LinkedIn and YouTube posts are named without their numbers). It also posts the day's research picks, with their sources and the @handles we credit, and the items it turned down, with the reasons.
Retention and deletion
Post statistics are refreshed every 6 hours for 45 days after publishing (X posts once a day, for their first 14 days). After that, the last numbers stay with the post's record, with two exceptions. LinkedIn statistics are deleted 12 months after the post was published. YouTube statistics are refreshed every 6 hours for as long as the record exists, and if they can no longer be refreshed (the video was removed or access was revoked), we delete them within 30 days.
Post ideas stay in the CRM until we delete them. When an idea is saved, the titles and channel names of other channels' YouTube videos from that week's research are replaced with a general reference, and links to those videos are removed from ideas, in their links and their text, 28 days after the idea was created. The list of X accounts the tool has looked up is kept until we clear it, and the list of sources it turned down for 14 days. The weekly summary is replaced each week. Messages the tool has posted in Slack stay in our workspace's history; they carry no LinkedIn or YouTube statistics.
The CRM also keeps a change history of each record. Statistics leave a post's change history after a day, and links to other channels' YouTube videos leave an idea's change history 28 days after the change.
Permanently deleting a post's record in the CRM removes its stored IDs and statistics. Copies in our backups expire on their own: nightly CRM backups in Amazon S3 after 35 days, and daily server disk snapshots after 7 days. To ask us to delete data, email privacy@apptycoons.com; we act on requests within 30 days.
Revoking access
The owner or admin of each account can cut off the tool's access at any time:
- YouTube: the account owner can revoke the tool's access to the YouTube channel from Google's security settings: https://security.google.com/settings/security/permissions.
- LinkedIn: the page admin who connected the tool can remove it under Settings & Privacy > Data privacy > Other applications > Permitted services: https://www.linkedin.com/psettings/permitted-services.
- Facebook and Instagram: an admin of our Meta business portfolio can revoke the tool's token, or remove the Page and Instagram account from it, in Meta's Business settings under Users > System users.
- X: the account owner can revoke the app in X's settings under Security and account access > Apps and sessions > Connected apps.
When access is revoked, the tool stops publishing and reading statistics on that platform, and we delete its tokens from Parameter Store, the server's configuration and the tool's data folder. A revoked token no longer works, and disk snapshots that still hold a copy expire within 7 days.
Contact
Apptycoons LLC, 8403 Pines Blvd, Ste 217, Pembroke Pines, FL 33024, USA · privacy@apptycoons.com
Our founder is responsible for privacy compliance and handles every request sent to that address.