ShelfLife Ember builds your Sponsored Products campaigns, adjusts bids daily against rules you can read, and shows you every change it made in plain English.
Waitlist members get founding-member pricing and priority onboarding. Ad spend is paid to Amazon separately.
Ad spend, 30d
498 clicks at $1.16 average
Ad sales, 30d
100 orders through ads
ACOS
Ad spend divided by ad sales
Cost per reader
Ad spend divided by orders
Illustrative figures for ShelfLife's own book. Amazon reports ad data on a delay, so the most recent days always under-report.
Ad spend, 30d
498 clicks at $1.16 average
Ad sales, 30d
100 orders through ads
ACOS
Ad spend divided by ad sales
Conversion rate
Orders divided by clicks
Cost per reader
Ad spend divided by orders
Cost per reader is ad spend divided by orders for the period shown. Falling is good, because each new reader is costing less.
Orders placed through ads for the period shown, in whichever view is selected above.
Illustrative figures for ShelfLife's own book.
Most ad platforms show you a number and leave you to guess how it got there. These are the panels an Ember author opens.
Ember reviews your targeting once a day. Anything it changes gets logged with the reason and the keyword, so you can read a week of activity in about a minute.
It also records the evaluations that led nowhere. When Ember looks at 1,847 targets and leaves 1,806 of them alone, that is written down too, along with the numbers it compared. Inaction is a decision, and you get to audit it.
What Ember did
The last 7 days. Ember reviews this book's targeting every day.
Aug 14
Lowered bids on 5 targets that kept spending without selling: "book marketing for authors", "how to sell more books" and 3 more.
Paused "publishing companies". Clicks kept coming without an order.
Aug 13
Promoted 2 converting searches into their own keywords so their bids can be tuned individually: "book launch checklist", "amazon ads for authors".
Aug 12
Raised bids on 23 targets that were not showing up in enough searches.
Lowered bids on 3 targets that kept spending without selling.
Aug 11
Reviewed 1,847 targets and changed nothing on 1,806 of them. Each one is logged with the numbers it compared.
Plus 94 more changes in this period.
Budget protected is the money that went to targeting Ember has since switched off. It is the clearest measure of whether the pruning is working, and it goes on your dashboard whether the figure flatters us or not.
Ember never touches your daily budget on its own. It moves bids and it turns targeting off. Your ceiling stays where you set it.
Budget protected, 30 days
Spent in the last 30 days on 9 targets Ember has since paused. Ember prunes targeting that spends without selling.
Who's finding this book
| Theme | Impr. | Orders |
|---|---|---|
| Self publishing & KDP | 61.4% | 78.5% |
| Book marketing & promotion | 24.1% | 16.9% |
| Author business & platform | 8.7% | 4.6% |
| Writing craft | 5.2% | 0.0% |
| Everything else | 0.6% | 0.0% |
Self publishing and KDP draws 61.4% of impressions but 78.5% of orders, so it converts well above its share. Writing craft has real reach and no orders yet.
Ember groups the search terms that found your book into concepts, then scores each one on how well it converts and how many orders it has actually produced. A concept with three clicks and one lucky sale does not get to sit at the top of the list.
Authors have told us this panel is worth the subscription on its own, separately from the ads. It is the closest thing to a straight answer about what your readers were looking for when they found you.
Top concepts
| Concept | Impr. | Clicks | CVR | Score |
|---|---|---|---|---|
| book marketing · 16 searches | 14.2k | 118 | 13.6% | 100 |
| self publishing books · 11 searches | 9.6k | 74 | 9.5% | 61 |
| how to market a book · 9 searches | 5.1k | 46 | 10.9% | 58 |
| author platform · 7 searches | 2.4k | 21 | 9.5% | 37 |
| book launch strategy · 12 searches | 1.9k | 18 | 11.1% | 34 |
Ember is not a dashboard sitting on top of Amazon's own automated bidding, and it is not a language model asked to make judgment calls with your budget. It is a system our team wrote, running logic our team decided on, built from over a century of combined Amazon advertising experience.
Every threshold in the system came from watching what happened to real author campaigns across $10M in spend, not from Amazon's public best-practice documentation.
Given the same numbers, Ember reaches the same decision every time. Any decision it made can be replayed against your history and checked. Nothing drifts, nothing changes its mind overnight.
Every adjustment is logged with the reason and the numbers behind it, including the evaluations that ended in leaving a target alone.
Each category of change has its own setting. Move bid changes to automatic once you trust them and keep pauses on approval, or run the whole thing on approval until you have seen a month of decisions.
Ember makes the change during its daily run and logs it. You read about it afterward.
Changes collect into a batch you approve or reject together. Nothing happens to your account until you say so.
Ember still evaluates and still logs, but a human reviews before anything reaches Amazon.
Two safeguards sit on top of your setting. Each category has a dollar cap that trims how large a single change can be. And every change gets scored for risk before it runs, using the money at stake, how easily it could be undone, and how much data supports it. A change that scores too high gets held for your approval even if you set that category to automatic. Risk can move a decision toward asking you. It can never move one away.
Setup takes real input from you. We would rather say that plainly than promise a two-click launch and have you discover otherwise.
ASIN, monthly budget, seed keywords and competitor titles to target, and your manuscript file.
Add our email as an admin inside your own Amazon Ads console. Your account stays yours and you can revoke it whenever you want.
Our team checks your listing is eligible and your seeds are sound, then launches eight campaigns. If anything fails the check we stop and tell you rather than launching something half-built.
Ember pulls yesterday's numbers, evaluates every target, makes its changes and writes them down. Repeat.
Every threshold on this page came from watching what happened to real campaigns across six years and hundreds of author accounts.
I've worked with more agencies than I'd like to admit because most of them are rarely ever able to produce the results they promise. Alex and his team are the first agency out of every single one I've worked with that has truly under-promised and over-delivered. Within a few months, they have tripled the results of my previous agencies and what I have been able to do myself.
Every subscriber gets the full rule set, the full dashboard, and the same team that runs our agency accounts.
The $189 is the platform. Your ad spend is separate and paid directly to Amazon on your own card.
We suggest starting at $500 a month in ad spend. Below that, campaigns rarely gather enough clicks and orders for the rules to have anything to work with, and you end up paying discovery prices without reaching the point where the tuning pays off.
Nobody will force you to spend it. But if the budget is going to be small, the honest advice is to wait until it isn't.
Publishing a second title? Each book runs as its own subscription, because each book gets its own campaigns, its own budget and its own audience data.
Reading this before you join saves us both a refund conversation.
The dashboard figures on this page come from Before the Bestseller, which is Alex's book, running on the same system you would be signing up for. When a rule misfires we find out on our own account before anyone else does.
For six years our team has run Amazon Ads for some of the best selling authors in the world. Ember is that work written down as rules and priced so an author with one title can run it. We kept the thresholds public because the thing authors have been burned by is not bad software. It is software that would not tell them what it did.
Join the waitlistEach book. One subscription per title, because each book gets its own campaign structure, its own budget, its own targeting and its own reporting. Two books running at once is $378 a month plus whatever you spend on Amazon.
The text of the book feeds the targeting work. It informs which keywords are worth seeding, how search terms get grouped into concepts, and which competitor titles are genuinely adjacent rather than superficially similar. A cover and a blurb tell us what the book is marketed as. The manuscript tells us what it is.
The file is encrypted, it is never used for anything besides your own campaigns, and it is deleted 60 days after your subscription ends. Pausing does not start that clock.
No, though you will end up understanding it better than you expected, because every change comes with its reason attached. The setup asks for seed keywords and competitor titles, and our team helps if you are not sure what to put there.
The part that spends your money is not. Bids, pauses and harvesting run on fixed rules with published thresholds, which is why we can print the whole rule set on this page. We do use a language model for one narrow job, grouping search terms into readable themes for the concept panel, and that step cannot change a bid or spend a dollar.
Once a day. Amazon delivers ad reports on a delay and attributes orders for up to 14 days after the click, so the most recent days always look worse than they turn out to be. Ember deliberately evaluates against data that has settled rather than reacting to a partial day, which is a limitation we would rather state than paper over with a live-looking dashboard.
That is a normal starting point and the early weeks look different because of it. With no order history the rules spend their first stretch finding out which searches convert at all, so expect discovery numbers before you expect efficient ones. The activity log makes that phase legible instead of alarming.
Only in a category you set to automatic, within the dollar cap you set for it, and never on your own author name. Anything scoring above your risk threshold waits for your approval regardless. Every change is reversible and every change is logged, so you can find it and undo it.
The agency is custom work for authors and publishers running large launches, with strategy, campaign types beyond Sponsored Products, and a person assigned to the account. Ember is the Sponsored Products playbook from that work, automated and priced per book.
Anytime, from your dashboard. No contract, no annual commitment, no setup fee. You can also pause advertising without pausing billing, or pause billing and keep the ads running through the period you already paid for. Those are separate switches on purpose.
We are letting waitlist members in first, in small groups, so the support stays real. Join and we will be in touch with timing and your founding-member price.
Founding members get early access pricing, priority onboarding, and a direct line to the ShelfLife team.
$189 per book per month at launch. $500 a month in ad spend recommended. Cancel anytime.
