Compare
Poool is a hosted script your revenue team drives from a dashboard. EdgeKit is a global CDN in front of your site, with your reader record in a database you can name. They differ on whether the paid text ever leaves your server, who holds the reader data, and whether a price exists before a sales call.
10 questions · checked 4 September 2026
Side by side
Rows are ranked by how much each one moves a decision. Six come out even, and the four EdgeKit takes are all about where things physically live.
| The question | Poool | EdgeKit |
|---|---|---|
| Does the paid text ever reach the reader’s browser?Favours EdgeKit | Three documented locking modes: hide the text with CSS, trim it in the browser, or take an event and write the enforcement yourself. In the first two it was delivered. | No. EdgeKit’s global CDN sits in front of your own site, decides server-side and withholds the gated body, so the paid text never leaves your origin. It asks one thing of your template: a CSS selector naming the article body. Half the paywalled sites we surveyed already declared one; the rest add it in a single deploy, and we verify it per title before you sign. |
| Can we run our own SQL against our reader data?Favours EdgeKit | Reader data sits in Poool’s platform. Flow, described as their data export solution, is the documented way to get a copy of it back out. | Your reader record is a database of your own. Ask for managed deployment and that database is a project in your own cloud account. |
| Can we get a price before we talk to anyone?Favours EdgeKit | No published price and no pricing page. The route to a number is a demo booking or a sales email; no figure appears on the site. | The whole rate card is on the site: eight meters, an included amount and a rate on each, and a $99 monthly floor. |
| Does the paywall learn on its own, or do we write every rule?Neither | An engagement score computed in their dashboard from recency, frequency and pages per session, 0 to 100, sorting readers into volatile, occasional, regular and fan, each with its own journey. It is a marketing formula rather than a model, the same shape EdgeKit computes; machine-learned propensity arrives through a partner, Sophi.io, whose score Poool takes in to pick the journey. | Rules you write, over scores EdgeKit already computes: Profile Streaming derives engagement, recency and frequency on the live event stream as it lands on the reader record, and a rule reads them in the same request — as it reads a score written back from a model you already run. What is not shipped is the model that learns the score: EdgeKit AI Paywall, scores an authored rule reads or the rules and the tests handed to the model to optimise for yield, is coming; ask us where it is. |
| How quickly can we be live?Neither | A hosted script and a dashboard. Their own published figure is integrated in under two days and live in a month, across 250+ publishers. | The same shape, if you want it: a browser script on the pages you already render, walls drawn client-side, nothing in front of your site and no traffic change. The CDN can come later. |
| Who changes the paywall, the developers or the revenue team?Neither | The revenue team. A journey builder, where a scenario is an ordered list of widgets the reader advances by completing the previous one, with scored reader groups behind it. | The revenue team. An ordered rule they write and deploy themselves, landing as one numbered entry naming who shipped it. |
| Can our data team reach the raw behavioural rows?Neither | Audit’s dashboards, plus Flow, their data export solution, for a copy of what sits underneath them. | Every event, unsampled, as Parquet in object storage you own, and the export is self-serve: CSV on demand from the console, or a scheduled export delivered to a bucket you name. |
| Can we keep the billing vendor we already run?Neither | Subscribe handles checkout and subscriptions, and their docs state Poool deliberately does not hold the identifiers for server-side Stripe payment, leaving those with you. | EdgeKit holds one billing model that every processor maps into — accounts, catalog, subscriptions and their items, schedules, invoices, payments, refunds and disputes — and every row records which system it came from. The paywall decision never branches on the provider, so connecting one EdgeKit has not connected before is a connector into a model that already exists, not a second billing model. Recurly and Stripe map into it today, each a full mirror kept current by webhook. Any other processor is wired by EdgeKit, at no charge, inside the migration rather than as a change order after it — and a group mid-cutover can run two at once. |
| We publish several titles. Is that one reader or several?Favours EdgeKit | Access and Subscribe are each started with an app id. Darwin CX, which agreed to acquire Poool in March 2026, brings a central subscriber database behind them. | Your choice, separate or combined: one reader record behind every title with one agreement and one set of counts, or a separate record per title. |
| Who splits the traffic for a test, and who reads the result?Neither | Poool. A test is a second scenario in the same context, with a slider setting the split between them, and it runs two versions at a time. Results are the two scenarios drawn against each other in the dashboard’s analysis view. | EdgeKit, or the experimentation vendor you already run. Either EdgeKit assigns the arm itself, or it reads the one your vendor assigned and decides on that — so the split and the statistics can stay in the tool your team already uses. |
What decides it
A hosted script can only act on a page the browser already has. Poool’s own documentation says so: hide it with CSS, trim it in the browser, or take the event and do it yourself. For a great many publishers that is the right trade, because it is fast, cacheable and survives a CMS nobody wants to touch. EdgeKit puts its global CDN in front of the site instead: the decision runs before the response is written and the gated body is never in it.
EdgeKit provisions one database per media group at signup, and if you ask for managed deployment it is a project in your own cloud account with EdgeKit operating it on your infrastructure. Neither is an upgrade tier. Where the reader record lives is your decision.
Eight meters, an included amount and a rate on each, a $99 monthly floor, a calculator that asks for no email, and the point at which the published list stops and a quote begins. Poool publishes no figure at all, so the comparison is between modelling your own bill and booking a call to find out.
Questions
The definitive agreement was announced on 4 March 2026 from Toronto and Paris, price undisclosed, with Poool’s Bordeaux team staying in place and existing client setups unchanged. We have found no published closing announcement, so the agreement is the last public step and we will not characterise the deal as completed. Everything this page says about Poool is checked against Poool’s own live documentation, which is the product a buyer would be signing for today. Ask them about the pairing: Darwin CX brings subscription management, authentication and a central subscriber database behind Poool’s front end.
In two of their three documented modes, yes, and Poool’s own documentation says so. Their configuration page describes hide, which leaves the content where it is and uses CSS to conceal a percentage of it, and excerpt, which moves the original content into a holder and replaces it with trimmed text. The third mode, custom, does nothing to your content and fires events so you can write the enforcement yourself. That is the one mode where a Poool implementation can withhold at the server, with your code doing the withholding. That is the normal trade for a client-side paywall, not a fault in the product, and it is the row where the two architectures differ most.
We can only answer half of it. EdgeKit publishes every rate: eight meters, an included amount and a per-unit rate on each, and a $99 monthly floor, with a calculator you drive with your own volumes. Poool publishes no figure and has no pricing page. The route to a number is a demo booking. We have seen no primary source for any Poool contract value and we do not repeat the numbers that circulate. Ask them for a per-unit basis and an annual uplift clause in writing, then run our calculator on your own volumes.
The data is yours and there is no export fee, here or in the contract. Export is self-serve: CSV on demand from the console, or a scheduled export delivered to your own object storage, beside the event stream that already lands there continuously. Where EdgeKit is deployed into your own cloud account, the database and the event archive sit in infrastructure you control and you do not need us to reach them at all. Poool ships Flow as its documented data export solution; on this question the two are level.
Show you 250 publishers already running it. Poool has been shipping since 2016 and names its clients; every row on this page is about the product rather than the track record, and you should weigh the track record separately. On the product itself the gap is the learned model: both platforms score readers on how recently and how often they read, and Poool additionally takes a partner’s machine-learned propensity score in, where EdgeKit’s own AI Paywall is coming rather than shipped.
How we checked
If a claim here is wrong or out of date, write toinfo@edgekit.cloud and we will correct the page and say what changed. That includes anybody at Poool, which Darwin CX has agreed to acquire.
Poool and Darwin CX are trademarks of their respective owners. EdgeKit is not affiliated with, endorsed by or sponsored by either. Product names are used here only to identify the products being compared.
Next step
A walkthrough against your actual setup, run by somebody who has run a subscription media business, including the parts of this page where Poool wins.
Book a demo