The marketing tools American teams actually use day to day, and the ones to cut this quarter
The eight categories a real American stack contains, the four that get bought and abandoned, and the test that tells which of yours is which.
Published by AI Knows Us · Updated 22 October 2026
A working American marketing stack is smaller than the logo slide suggests. Eight categories get opened most days. Four more get bought, demonstrated internally once, and then quietly renewed for two years. The test that separates them is not usage analytics, which overstates logins, but a question: name the decision this tool changed in the last ninety days. A tool that cannot have that sentence written about it is a subscription rather than a tool, and cutting it funds the thing you actually need. Here are the eight, the four, and the arithmetic for the one category most American teams are now adding.
The eight categories opened most days
| Category | Why it survives | The decision it changes |
|---|---|---|
| Web analytics | It is the shared factual base for every argument | Where to put next month's effort |
| CRM | It holds the pipeline and the source field | Which channel is producing revenue rather than leads |
| Email and lifecycle platform | It is the one channel you own outright | What to send, to whom, when |
| Ad platform native interfaces | Money moves here daily | Bids, budgets, negatives |
| Call tracking | In most American service categories the phone is the conversion | Whether a campaign is working at all |
| Search console and keyword data | It reports the queries you actually appear for | What to write and what to fix |
| A design tool | Assets are needed constantly and briefs are slow | Whether a campaign ships this week |
| Project or task tracking | It is how a small team stays sequenced | What gets dropped when the week shortens |
The four that get bought and abandoned
- Multi touch attribution software, below a few hundred steady conversions a month. The output looks authoritative at any volume and is only meaningful above it.
- A customer data platform, with fewer than a dozen systems holding customer records. The benefit at smaller scale is a clean CRM, which costs a fraction as much.
- An enterprise content management replatform, without hundreds of editors or several languages. Fixing the existing templates and internal linking delivers most of the gain.
- Social listening, bought as a strategy input. It works as an alerting tool and it rarely survives as a planning input, because nobody owns the weekly read.
In all four cases the failure is ownership rather than capability. A platform with no named operator by month four becomes a renewal, and the renewal is easier to approve than the cancellation.
The American cost lines that change the real number
Three, and all three are routinely left out of a stack budget. Implementation is often a multiple of the first year licence and is quoted separately or not at all. The next seat matters more on an American marketing team than the headline volume allowance does, so ask the price of seat eleven rather than seat ten. And sales tax is a state question rather than a price question, because software subscriptions are taxed where the customer is and whether a particular subscription is taxable differs from state to state. Two American companies buying the identical stack on the same day can receive different invoices. Ask each vendor how it treats your state and write the answer in the row beside the price.
The ninth category most American teams are now adding, and how to size it
Reading what the assistants say is the newest line in an American stack, and it is the one most likely to be bought at the wrong size. Do the arithmetic before you look at a plan. Tracked buyer questions, times engines read, times refresh frequency, times metros or markets. Forty questions on five engines read weekly is 800 recorded answers a month for one market, and the same forty across four metros is 160 tracked questions and 3,200 recorded answers. Vendors meter differently, so ask which unit the price is per: a tracked question, an engine, a recorded answer or a domain. Then recompute your total before comparing two quotes, because two honest quotes for identical work can differ by an order of magnitude on the meter alone.
What that category reads is worth knowing before you buy it. On 22 October 2026 we put a 198 question American buyer bank to the ChatGPT and Gemini product surfaces through a data provider, with the United States named inside each question, because the provider accepts gl, hl and location parameters and ignores all three. The brand batch is excluded as a ceiling measurement. In 64 blind American readings the engines produced 2,866 citations across 1,143 distinct hosts, named a rival in 24 readings, named nobody in 38, and named us zero times. All 48 supplier namings went to software products, led by Semrush at 14, HubSpot at 12, Ahrefs at 9 and BrightLocal at 5. ChatGPT ran a live search on all 32 of its American readings and Gemini on 16 of its 32, so a tool reporting one blended score is averaging two instruments with different refresh behaviour, and the engine split is a feature to ask for rather than a nicety.
What to do at the next renewal
- Write the sentence for every tool. The decision this changed in the last ninety days. No sentence, no renewal.
- Name an operator for every tool you keep. A tool without one is already a subscription.
- Ask every vendor the price of the next seat and the treatment of your state's sales tax, and record both.
- Size the assistant reading with the multiplication rather than by plan name, and ask for the engine split explicitly.
- Put the money saved from the cuts into the two categories that reliably earn, which in most American stacks are call tracking and the email platform.
If the category you are adding is the assistant reading, the per unit and per location traps are costed in full in why the unit count multiplies before the dollar figure does. If what you need is a free tier rather than a purchase, where each one stops is audited in where a free tier stops.
Questions about American marketing tool stacks
How many tools should a ten person American team have?
Roughly the eight categories above, which is usually six to nine products once some overlap. Beyond that, each addition needs an operator and a decision it changes.
Is call tracking really necessary?
In most American service categories yes, because the phone is where the conversion happens and an untracked call makes a working campaign look like a failure.
What is the fastest cut to make?
Anything bought for a capability nobody owns. Social listening and attribution software are the two most commonly in that position at mid size.
Should the stack include an AI visibility tool?
If buyers in your category ask assistants, yes, sized by the multiplication above rather than by the plan name, and chosen partly on whether it reports the engines separately.
Why does the engine split matter in a tool choice?
Because on our own reading one engine read the live web on every American question and the other on half of them, so a blended score would have hidden the finding most relevant to how fast new publishing takes effect.
The newest line in the stack, and what to demand of it
Of the nine categories above, the assistant reading is the one where a buyer is most likely to be sold a score with nothing under it. AI Knows Us reads ChatGPT, Gemini, Perplexity, Claude and Google AI Overviews for the buying questions in a category, reports each engine separately with its denominator, and keeps the raw answer text so a figure in a monthly pack can be audited rather than accepted. In the United States that is the Growth plan at 299 dollars a month with 100 buyer questions watched and 1,600 credits, read on aiknowsus.com/pricing and aiknowsus.com/llms.txt on 22 October 2026, with credit top ups at 49 dollars per 1,000.