You now have a list of keywords where your product already beats the category average. The next question is which to attack first — and the answer is not "all of them".
Keywords are not independent. Ranking for one lifts related ones, which means the right unit of work is a group, not a keyword.
Prefer to watch? One hundred and fifteen seconds on why the sales signal transfers across related terms, how to build a root-and-long-tail group, and how each group won frees the budget.
Why groups beat individual keywords
Amazon promotes products it believes are best, and that judgement generalises across closely related terms.
If your standing desk converts well and sells heavily for standing desk, your rank improves not only there but for standing desk for office and wooden standing desk too. The sales signal transfers.
So every sale from any keyword in a group helps every other keyword in that group. Targeting a coherent group concentrates the benefit; targeting scattered unrelated keywords spreads it until none of it registers.
Building the master list
Start from the keywords your research marked as worth targeting — the ones where your CVR beats the keyword average.
Take them from both views. Some keywords will qualify on organic SQP data, others only on advertising data, and the sets do not fully overlap. Merge them, deduplicated.
Then reconsider the near-misses. Look at the keywords you set aside, and specifically at large keywords where your CVR is close to the average rather than clearly below it.
These are judgement calls. Their odds are lower, but a big keyword you are nearly competitive on can be worth testing — especially if the launch tactics in the next stage will push you over the line. Add them deliberately, knowing they are the speculative part of the list.
Root keywords and long-tail keywords
Each group is built from one root and its long tails.
Root keyword — one to three words, the general term. High search volume, high sales volume, high competition. Bamboo bed sheets.
Long-tail keywords — three or more words containing the root. Lower volume, but usually higher purchase intent and better conversion, because the shopper has told you exactly what they want.
| Root | Long-tail |
|---|---|
| Bamboo bed sheets | king size bamboo bed sheets |
| bamboo bed sheets for women | |
| pink bamboo sheets for queen size bed | |
| Protein snack | chocolate protein snack |
| protein snack big pack | |
| 25g protein snacks for the gym |
The root is what you are really trying to rank for. The long tails are easier wins that feed the same signal — they convert better, so they help your averages while contributing sales that lift the root.
Prioritising the groups
Rank your groups and start with one. Two at most.
What makes a group a good first target:
Your CVR and CTR clearly beat the average on the root and on several long tails. Strongest possible evidence you can win it.
The root has real purchase volume. Search volume without purchases is a group that cannot repay the spend.
Your current organic rank leaves room. Already at position 8 means limited upside; position 60 with strong conversion metrics is the ideal profile.
Purchase intent is high across the group. Prefer groups where the terms describe a product rather than a situation.
The long tails are genuinely related. A group of terms that only share a word but not a shopper intent will not transfer signal.
One group at a time. Genuinely.
This is where most launches fail, and it is worth stating plainly.
To rank for a keyword you must be among the products driving the most sales for it. That is a relative threshold, not an absolute one.
Splitting a limited budget across five groups means falling short of the threshold in all five. You will have spent the entire budget and moved nothing. The same money on one group can carry it past the threshold and into a rank that then pays for the next group.
Half the required velocity does not get you half the rank. It gets you nothing.
Then work down the list
Once a group reaches the best rank it is going to reach, and holds it, move to the next.
That sequencing is the whole long-term plan: group by group, the product accumulates organic positions, each one reducing how much advertising the previous group still needs. Over time the account converts paid velocity into free traffic.
Tracking the group
From here you need rank on every keyword in the group, checked regularly, plus their CVR and CTR against the averages.
TrackIQ's rank tracking and Search Query Performance views cover both over a consistent date range, which matters here because you will be checking the same terms repeatedly for weeks and need the comparison to be like-for-like. See also Triage a Rank Drop for reading rank movement that is not caused by your launch.
Common mistakes
Grouping by word rather than by intent. Standing desk and standing desk mat share words and not shoppers.
Starting with the biggest root keyword regardless of whether your metrics support it.
Running four groups because the budget technically allows it.
Ignoring long tails because the volume looks trivial. They convert better and they feed the root.
Related
- Keyword Research from Search Query Performance — producing the list this groups.
- Preparing a Launch: CTR and CVR Tactics — what to do before spending anything.
- What a Relaunch Is, and When to Run One — why velocity is a relative measure.