Answers come from the product guides and the live pricing
Ask me what the platform does, what it costs, or how a part of it works. I answer from the published guides and the current plans, and I say so when something is not in them.
Try one of these
Written by an assistant. Nothing here is a quote or a contract.
How to use
Marketing Management
Checked against the product on 27 August 2026
Marketing Management
Marketing covers brand, content, social, search, advertising, reviews, creators and budget. It is a paid App.
One principle runs through the whole thing and explains most of its behaviour:
This App proposes, records and reports. It does not spend your money and it does not move your budgets. Nothing here can pause a campaign or change an ad budget. Those happen in the ad network, by a person.
Every page has a How to use button in its header.
Where everything lives
Group
Pages
Director
Recommendations, Decisions, Brands Budget
Brand
Brands, Brand Kit, Media Library
Content
Content Studio, Calendar, Image Studio, Video Studio
Social
Networks, Posts, Inbox, Reviews, Post viewer, EGC Agent, My Advocacy
Search
Site Audit, Rank Tracking, AI Visibility, Agent Readiness, Agent Surface
Advertising
Ad Accounts, Spend and Return, Insights, Creators, Collaborations
Capture
Landing Pages, Forms, Subscriptions
Audience
People
Automation
Journeys
Settings
Marketing Settings
Brands are the organising idea. Most screens are scoped to the brand chosen at the top of the page, end to end.
The Director
The Director reads your brand kit, your pipeline, your ad spend and your content calendar, then writes a short read of where you are plus a handful of specific recommendations. Because it can see your deals, it can say which lead sources actually close rather than which ones produce the most clicks.
The rules it works under
It proposes. You decide. Nothing runs by itself. Agreeing with a recommendation records your name and the date against it. It does not move a budget, pause a campaign or publish anything.
Agreed is not done. Mark something done once you have actually done it, wherever that was. Keeping the two apart is what stops a list of intentions being read later as a list of outcomes.
Every recommendation cites real figures, frozen at the moment the plan was written, so a proposal you read next month still shows the numbers it was argued from rather than quietly re deriving new ones.
What it could not see
Skipped sections are printed above the plan. If ad spend was not readable, the plan says so before it says anything else. A plan that quietly analysed four of five sections reads exactly like a complete one.
A locked section shows the reason, usually a module that is off or a permission you do not hold. Reading the pipeline needs permission to see every lead in the workspace, not only your own, because a plan computed from one person's leads would be a fragment presented with the confidence of the whole.
Asking for a plan costs an AI call and reads a lot of data, so it is a separate permission from reading one. Every plan is kept, including the ones that failed.
Social
The Networks page is the reference the composer is built on. Every rule shown there is the same rule that checks a post before it is saved, so nothing on that screen can promise something the publish step will later refuse.
Read it before you write, not after
Each card leads with what the network cannot do. LinkedIn has no way to post a poll. Instagram has no text only post at all. TikTok will not let a post be deleted once it has gone out. Finding that out after the words are written and the picture is made leaves you nothing to do but start again.
Every refusal says whose rule it is. Almost all of them are the network's own limits, not ours.
A network being listed does not mean it is connected.
Things that catch people out
Six of the twelve networks need your own developer app. LinkedIn, X, YouTube, TikTok, Pinterest and Google Business Profile all require your company to register and be approved itself. That review is measured in weeks, so start it before you plan a launch around it.
TikTok keeps posts made through its connection private until its audit passes. Everything published that way is visible only to the account owner until then. It catches everybody once.
Two of them are not a login at all. Bluesky uses an app password you generate in your own settings, and Telegram uses a bot you create and add to your channel. Neither has a consent screen.
There are hard daily caps. Instagram allows 25 posts per account in any rolling 24 hours. YouTube allows roughly six uploads a day. Threads allows 250. When a post is held back, the queue tells you which limit it was and when it clears.
Half of them cannot be edited after publishing. Fixing a typo means deleting and reposting, which loses the likes and comments the original had.
How limits and timing work
A rolling window, not a calendar day. Instagram's 25 posts do not reset at midnight. The allowance comes back as your oldest post ages out.
Most networks will not hold a post for you. Only Facebook, YouTube and Mastodon accept a post now and publish it later. For the other nine the schedule is kept here and the post is sent at the moment it is due, which is why the service has to be running for a scheduled post to go out.
Character counts are not the same everywhere. X counts every link as 23 characters however long it really is. The composer shows the count each network will actually see.
Mastodon is per server. Every server sets its own limits. The 500 characters on its card is the common default, not a rule: when you connect an account the real numbers are read from your own server.
These numbers are a snapshot of each network's published documentation, with the date it was taken shown on every card. If something disagrees with what a network tells you, the network is right.
Search
Three screens with three different jobs:
Site Audit crawls your website and scores it against a published checklist, then gives you a fix list in the order somebody would actually work through it.
Rank Tracking follows where you appear for phrases you choose.
AI Visibility asks an assistant the questions a buyer would ask and records whether you come up.
The score is an opinion, and you can read it
Nobody outside a search engine knows what ranks. What can honestly be said is how your site does against a published checklist. Every check, what it measures and what it is worth is behind "read the checklist" on the audit screen. If you disagree with a weight, you can see the weight.
A score only compares with another score from the same checklist version. When the checklist changes, older scores are marked as not comparable rather than quietly charted beside new ones.
What the audit does to your website
It reads pages. It changes nothing, ever. It requests one page at a time with a pause between them, and it obeys your robots.txt even though the site is yours: a page you have disallowed is reported as skipped rather than fetched anyway.
A page that could not be read is listed with the reason rather than dropped, because a page quietly missing from a report reads as a page that passed.
"I fixed this" is a note, not a result
Marking something fixed records that you say you fixed it. Only the next audit can confirm it, and until then the item says so. When a later audit no longer finds the problem it is marked confirmed. Two different facts, two different words.
Rank data is bought, not scraped
There is no lawful and reliable way to read a search engine at scale, so it is not done. You bring a key from a search data provider, exactly as you bring your own AI key. You pay them directly, nothing is charged for it here, and your bill is never seen. Add the key in Marketing Settings.
"Not in the top 100" is not a position of zero and is never charted as one. The country and the device are part of what a position means, so the same phrase in two countries is two rows with two histories.
AI visibility is a sample, and a small one
A run asks up to eight questions once, on one day, of one model. Answers vary between runs and between models, so two runs that disagree are normal.
It is not a share of voice and it is not a ranking, which is why there is no trend chart: a line through samples of size eight would be a picture of noise.
"Mentioned" means one of your brand terms appeared in the answer. It does not mean recommended, and it will count an answer that says something unflattering. Every answer is stored so you can read what was actually said.
The AI crawler list
Your robots.txt either lets each published AI crawler in or it does not, and that has a direct consequence: a blocked crawler cannot cite you. Blocking them is a perfectly legitimate choice. What is worth catching is blocking them by accident, because the line arrived in a template somebody installed.
What is generated, and what is deliberately left blank
FAQ structured data is built from questions already on your page, with the answers left empty for you to write. They could be filled from a model and deliberately are not: structured data saying something your page does not say is worse than none, and it would be published on your site under your name.
Advertising
This records what each campaign cost, day by day, and puts it beside the revenue that actually closed. It never creates, pauses or re budgets anything.
It works with nothing connected
You do not need to connect an ad account at all:
Press Track by hand on the Ad Accounts page. Choose the network, name it, and set the currency you are billed in. LinkedIn, TikTok, Microsoft, X, Pinterest, Reddit, Amazon and anything else are all available here, including the ones that cannot be connected.
Add a campaign and its destination. Type where the ad sends people and copy the tagged link it builds. Paste that into the ad, not your plain URL. The tag is the same one the report joins on, so the two cannot disagree.
Record what you spent: a date range and an amount, read off the network's own billing screen. A range is spread evenly across its days. Entering the same range again replaces it, so that is how you correct a figure rather than doubling it.
The return figure is the same one either way. Revenue comes from the tag on the link somebody clicked, joined to your own closed deals. No advertising connection is involved in that half, so a hand tracked campaign gets exactly the same answer as a connected one.
Connecting a network (Google and Meta only)
Add a credential. You register your own app with Google or Meta and paste its ID and secret. The drawer shows the exact redirect address to whitelist, and it has to match character for character or the consent screen refuses it.
Authorize it, signed in as somebody who can see the ad account. Read access only is requested.
Choose the accounts. One login often reaches several ad accounts, especially an agency login. Only the ones you pick are read.
It refreshes overnight, re reading the last seven days rather than only yesterday. Sync now does it immediately.
Two things that look like a fault and are not
Google: a new Developer Token sees nothing. It is issued with test account access only. It authorizes perfectly and returns no data for your real accounts until Google approves Basic Access, which is a separate application and usually takes a few days.
Meta: access stops after about 60 days. Meta issues nothing to renew in the background, so when it lapses you are told to click Authorize again. Your App ID and Secret are still correct and do not need re entering.
What the numbers mean
Spend is in the ad account's own currency, fixed by the network and often not the one your deals are booked in. Nothing converts or adds across currencies: two currencies get two totals, on purpose.
Yesterday's figure can still change. Both networks revise recent days for invalid clicks and late conversions, usually for up to three days. That is why a week is re read every night, and why a day that moved is marked as revised.
Conversions shown are the network's own claim. Where the metric is blank the account has no conversion tracking configured, which is not the same as no conversions.
Brands Budget
An answer to "what does our marketing actually cost per month".
Some of it is already measured: advertising spend, scheduled creator payments, and AI use against what the plan allows. The rest you enter yourself, because no integration can see your agency retainer.
It is scoped to the brand chosen at the top, end to end. What the platform charges you for the software is deliberately not here: that is a workspace cost that does not vary by brand, and it lives on the billing screen.
Setting what you plan to spend
A budget has two numbers: what you meant to spend, and what it is actually costing.
Set one with no brand selected and every brand inherits it, which is what a business that budgets centrally wants.
Set one on a brand and it overrides the shared figure for that currency.
A plan is a plan, not a cap. Passing it does not pause a campaign, block a collaboration or refuse an AI request. That is deliberate: acting on it would mean switching off your advertising from a figure typed on a screen, using spend numbers the ad networks keep revising for days afterwards.
Budgets are per currency. If you advertise in euros and pay an agency in dollars, set two. They are deliberately not added together: converting would need an exchange rate for the day of every transaction, and a total quietly wrong by an exchange rate is worse than several that are each right.
Your AI allowance is a burst limit, not a monthly quota
This is the one thing worth reading twice.
The AI request number on your plan applies to a rolling six hour window, not to a month. A plan showing 500 does not mean 500 requests a month. Spread evenly it is around 60,000.
That is why the screen shows two comparisons:
The monthly figure asks whether the amount of work fits at all.
The busiest day figure asks whether you are being turned away, and that is the one that predicts a refusal, because refusals are bursty.
You can sit at a small fraction of your monthly capacity and still be blocked every Monday morning.
The rest at a glance
Screen
What it is for
Brands
The brands everything else is scoped to
Brand Kit
Logos, colours, tone of voice, the material the AI reads
Media Library
Your images and files
Content Studio, Calendar
Planning and writing content, and seeing it on a calendar
Image Studio, Video Studio
Generating and editing visuals
Posts, Inbox, Post viewer
Scheduled and published posts, and replies coming back
Reviews
Reputation across review sites
EGC Agent, My Advocacy
Employee advocacy: colleagues sharing company content
Agent Readiness, Agent Surface
How well your site answers AI assistants
Creators, Collaborations
Working with external creators, and what you owe them
Landing Pages, Forms, Subscriptions
Capture, shared with the CRM App
People
Your marketing audience
Journeys
Automation, shared with the CRM App
Marketing Settings
Connections and keys, including your search data provider key
Billing and limits
Marketing Management is a paid App. Its price is shown on the Plan page.
Your plan caps AI requests, and as above that cap is a rolling six hour window, not a month.
Media you upload and images you generate count toward the workspace storage allowance shared across every App, shown on the Plan page.
Two costs are not ours and never appear on your bill here:
Your search data provider, for rank tracking. You bring the key and pay them directly.
Your own AI key, if you use one instead of the built in AI. Only built in AI is metered against your plan.
Frequently asked questions
Did the Director change my ad budget? No. Nothing in this App can. It proposes, you decide, and budgets move in the ad network by a person.
I agreed with a recommendation. Is it done? No. Agreed and done are separate on purpose. Mark it done once you have actually done it, wherever that was.
Why does my plan skip a section? Skipped sections are printed above the plan with the reason, usually a module that is off or a permission you do not hold. Reading the pipeline needs permission to see every lead in the workspace, not only your own.
Why do the figures in an old plan not match today's? They are frozen at the moment the plan was written, so the argument still shows the numbers it was made from.
Why can I not post a poll to LinkedIn? LinkedIn has no way to do it. The network card leads with what each network cannot do, and every refusal says whose rule it is.
My TikTok post is not visible to anybody. TikTok keeps posts made through its connection private until its audit of your app passes. That is TikTok, not a fault here.
Connecting a network is asking me to register a developer app. Six of the twelve require it: LinkedIn, X, YouTube, TikTok, Pinterest and Google Business Profile. That review takes weeks, so start it early.
My post was held back. You hit a network daily cap. Instagram allows 25 in a rolling 24 hours, YouTube about six, Threads 250. The queue names the limit and when it clears.
Instagram's limit did not reset at midnight. It is a rolling window. The allowance comes back as your oldest post ages out.
My scheduled post did not go out. Only Facebook, YouTube and Mastodon can be handed a post to publish later. For the other nine the schedule is held here and sent when due, so the service has to be running.
The character count differs between networks. X counts every link as 23 characters however long it is. The composer shows what each network will actually see. Mastodon limits are set by your own server.
I cannot edit a published post. Half the networks do not allow it. Deleting and reposting loses the original's likes and comments.
Why is my site audit score not comparable to last quarter's? The checklist changed. Scores only compare within one checklist version, and older ones are marked rather than charted beside new ones.
The audit skipped pages. Either your robots.txt disallows them, which is obeyed even on your own site, or they could not be read. Both are listed with the reason rather than dropped.
I marked an audit item fixed but it still shows. Marking it fixed records that you say so. Only the next audit can confirm it.
Why do I need my own key for rank tracking? There is no lawful, reliable way to read a search engine at scale. You bring a provider key, pay them directly, and nothing is charged for it here.
A keyword shows "not in the top 100". That is not a position of zero and is never charted as one. Country and device are part of what a position means.
Two AI visibility runs disagree. Normal. A run asks up to eight questions once, on one day, of one model. It is a sample, not a ranking, which is why there is no trend chart.
"Mentioned" sounds positive but the answer was not. Mentioned only means one of your brand terms appeared. It counts unflattering answers too. Read the stored answer.
Why are the FAQ answers blank? They are deliberately left for you to write. Structured data that says something your page does not say would be published on your site under your name.
Can I track an ad network you cannot connect to? Yes. Track by hand covers every network, and the return figure is calculated identically, because revenue comes from the link tag joined to your closed deals.
I entered the same date range twice. Did it double? No. Re entering a range replaces it. That is how you correct a figure.
My Google ads connection authorized but shows no data. A new Developer Token has test account access only until Google approves Basic Access, which is a separate application taking a few days.
My Meta connection stopped after two months. Meta issues nothing to renew in the background. Click Authorize again. Your App ID and Secret are still correct.
Yesterday's spend changed. Both networks revise recent days for up to about three days. A week is re read every night and changed days are marked revised.
Why are there two spend totals? You spend in two currencies. Nothing is converted or added across currencies, because a total quietly wrong by an exchange rate is worse than two that are each right.
I passed my budget and nothing stopped. A plan is a plan, not a cap. Acting on it would mean switching off advertising from a typed figure, using numbers the networks are still revising.
I am well under my monthly AI usage but keep getting refused. Your AI allowance is a rolling six hour burst limit, not a monthly quota. Read the busiest day figure, not the monthly one. That is the one that predicts refusals.
Something here not matching what you see on screen? The screen is right and the guide is what needs correcting. Tell us and we will fix it.
Marketing Management: how to use guide | ExecCortex
