Growth work fails in one of two ways.

You chase demand that was never there. Or you ship changes and cannot say whether any of them did anything.

This stack exists to close both gaps: the first three tools find and measure demand, the last three explain and keep it.

1. Semrush — find out what people are actually asking

Start with the market, not the idea.

What gets searched, how often, and by someone with a problem rather than idle curiosity. The best growth work is usually a page that answers a question people were already typing — you are finding demand, not creating it.

2. Ahrefs — find out whether you can win it

Volume without difficulty is a trap.

A term with ten thousand searches you cannot rank for is worth less than one with four hundred you can. Ahrefs is where you check who currently holds the answer and whether that is realistic to displace this quarter.

Be honest at this step. It saves months.

3. Google Analytics — know the shape of the funnel

Where people arrive, what they do next, and where they quietly leave.

GA is imperfect and it is the shared language — when someone asks whether traffic is up, this is the number they mean. Get the conversion events right and most reporting arguments disappear.

4. PostHog — find out whether the change worked

This is the step that separates growth work from redecorating.

Product analytics and experiments tell you whether the thing you shipped moved the number you claimed it would. Without it, every change is an opinion with a deploy attached.

Decide what would count as success before you ship. Afterwards is too late to be honest.

5. Hotjar — find out why

Analytics tells you that forty per cent leave on step two. It will never tell you why.

Watching ten sessions usually does, and usually in the first three. The answer is almost always duller and more fixable than the theory the team had.

6. Mailchimp — keep what you won

Traffic you have to buy again every month is not growth, it is rent.

An audience you can reach directly is the part that compounds. Put an email capture somewhere honest, send something worth opening, and stop paying to re-acquire the same people.

Do them in order

  • Demand first, so you are not optimising a page nobody wants
  • Difficulty next, so you pick fights you can win
  • Measurement before the change, not after
  • Qualitative last, to explain the number rather than replace it

Most growth stacks are strong at the top of this list and absent at the bottom. That is why so much growth work cannot be proved.

Do I need both Semrush and Ahrefs?
Not strictly — they overlap heavily. Teams that use both usually favour one for keyword and market research and the other for backlink and difficulty analysis. If you are choosing one, pick on the interface you will actually open.
Is Google Analytics still worth using?
Yes, as the shared baseline. It is the number other people mean when they ask about traffic. Pair it with product analytics for anything about behaviour inside the product, which is what GA is weakest at.
What does PostHog add over Google Analytics?
Event-level product analytics and experiments — whether a specific change moved a specific behaviour. GA is good at traffic and acquisition; PostHog is good at what happens after someone is already using the thing.
How many session replays should I watch?
Ten is usually enough to explain a drop-off, and the answer often appears in the first three. Replays are for explaining a number you already have, not for finding numbers.