Keap Review: Broad Capability, and the Hardest Exit We Have Measured

Keap sells one of the broadest products in this category and makes it the hardest to leave. Both halves of that sentence are measured, and holding them together is the only honest way to describe this product.

Start with what it is. Keap is a CRM with a marketing automation engine, an email and text platform, a landing page and form builder, a quoting and invoicing system, a checkout with order bumps and promotion codes, a native payment processor and an appointment scheduler, all inside one account^[2]. It was called Infusionsoft until 2019, and the old name still carries roughly a quarter of the search demand for the product. Since late 2024 it has belonged to Thryv, and the handover is visible in places a buyer should know about before signing.

OUR TAKE: The capability is real and we rate it near the top of the category. What pulls the score to 2.74, the lowest we have given any CRM, is not the software. It is that every movement which raises your invoice happens automatically, and every movement which lowers it requires a person on the telephone. The vendor publishes both rules itself. That asymmetry is the product's defining characteristic, and it is why the review platforms disagree about Keap by three full points.

## What you are actually buying

It is worth being concrete about the breadth, because it is the part that justifies the price when anything does.

On the customer record side you get contact management with company records, tags, saved searches and filters, custom fields, notes, tasks, lead scoring and a full automation history showing which sequences a contact has passed through and where they currently sit^[2]. On the marketing side, landing pages, multi-page funnels, forms, broadcast email, one-to-one email synced with Gmail or Outlook, newsletters, and two AI assistants that draft copy and assemble a campaign from a template. On the sales side, a pipeline, an appointment scheduler synced to your calendar, quotes that convert to invoices in a couple of clicks, recurring invoices, checkout forms, order bumps, promotion codes and a native payment processor.

Then the text and voice layer, which Keap calls its Business Line: one-to-one texts, automated texts, broadcasts, reply handling and a dedicated business number, with five hundred messages and a hundred voice minutes attached to every plan^[2]. Three different country footprints apply. The dedicated line and one-to-one texts cover the United States and Canada, text marketing is United States only, and the mobile app covers five countries.

The automation library is the piece buyers underrate. Keap ships prebuilt sequences in four groups, marketing, sales, service and operations, covering things like lead magnet delivery, flash sales, evergreen webinars, review requests, billing and new hire onboarding^[2]. Five of them are included on the entry plan, ten on the middle one, fifteen at the top. That count is the only capability number in the entire product that moves with price.

Put against the other six CRMs we have measured in this category, that span is first or second on every comparison we can draw from published feature lists. The depth inside each area is shallower than a specialist's: the pipeline is a pipeline rather than a forecasting system, and the reporting is adequate rather than strong. But a business that needs eight adequate things in one account, instead of four good products wired together, has very few real options at any price.

## One price on the page, ninety-nine in the payload

The pricing page shows a single figure: starting at $299 a month, billed monthly, beside a slider for contacts and users^[2]. It names no plan. We read the same page's data payload, which is the catalogue the vendor's billing system ships to every visitor's browser, and found 99 distinct priced items^[3].

Three of them are plans. They are called Ignite, Grow and Scale, those names appear nowhere in the rendered page, and they differ from one another in exactly three values.

Plan	Monthly	Annual, per month	Contacts	Users	Prebuilt automations
Ignite	$299	$249	1,500	2	5
Grow	$399	$329	2,500	3	10
Scale	$599	$499	5,000	5	15

That is the whole difference. Not one feature separates the cheapest plan from the dearest. The vendor says so on the page, in its own words: say goodbye to feature-based plans^[2]. It is telling the truth, and the consequence deserves stating plainly. A Keap plan is a meter reading. You cannot buy a cheaper one by giving up functionality, and you cannot buy a better product by paying more.

For a buyer, this is genuinely good news in one direction. Nobody on the entry plan is using a crippled version while the real product sits two tiers up, which is the normal arrangement in this category and the thing we criticise HubSpot and Salesforce for. On Keap the entry plan is the product.

The number that is not on the pricing page

Keap requires an implementation package. This is not our inference; the pricing page carries the phrase about required implementation services, and the pricing FAQ asks and answers why one is required^[2]. The answer is unusually candid for a vendor, and we will come back to it.

What the pricing page does not do is price it. The implementation services page shows one package at a struck-through fifteen hundred dollars reduced to five hundred^[7]. The catalogue carries implementation items at a thousand, two thousand and three thousand dollars, matched by name to the three plans, plus onboarding at nine hundred and ninety-nine and two thousand nine hundred and ninety-nine, plus an accelerator at five thousand^[3].

Both numbers are the vendor's. The page figure is a promotion and the catalogue is what the billing system holds. Take the catalogue figure matched to the entry plan, and a first year looks like this.

First year at the entry plan, in dollars	Amount
Twelve months of subscription, billed monthly	3,588
Implementation package matched to that plan	1,000
Total	4,588

The monthly price a buyer carries into a comparison is 78% of what the first year actually costs. The missing portion is the part the vendor itself calls required.

## The contacts you pay for are the most expensive contacts Keap sells

Here is a piece of arithmetic worth five minutes of anybody's time. The entry plan includes 1,500 contacts for $299 a month, which is just under twenty cents per contact. The first band of the add-on ladder sells a further thousand contacts for thirty-six dollars a month, which is three and a half cents each^[3].

The included contacts cost nearly six times what the additional ones cost. So the slider on the pricing page, which invites you to drag your contact count upward and watch the plan price rise, is steering you toward the worst rate in the catalogue. Size the plan down and the contacts up.

The ladder itself is reasonable: seventeen bands, from a thousand contacts at thirty-six dollars a month to more than four million at ten thousand one hundred and forty, with the per-contact rate falling the whole way^[3]. Scale is rewarded here. It is the one place in the product where it is.

## What a contact is, and why that matters more than it sounds

The vendor defines a contact as a single person recorded in your account^[2]. No exclusion is stated for duplicates, unsubscribes or hard bounces. Then two published rules meet, and the meeting is the single most consequential thing we found.

First, automated list management. A contact with no open, click or form submission for six months is moved to an unengaged, non-marketable status, and in the vendor's words such contacts are not eligible to receive your broadcast emails or automation emails^[9]. A warning status arrives at four months. The behaviour is on by default for all accounts, and a customer cannot manually reactivate a contact out of that status on any edition.

Second, the limits article. Going over your included contacts does not block anything: the vendor waits until your next billing date, raises your subscription's additional contacts and bills the increase^[10]. The same article answers the opposite question directly, and the answer is that Keap will not remove additional contacts from your subscription automatically.

Put them together. Records that have stopped being usable have not stopped being billable, and nothing reduces the bill except a deliberate act by a human. Whether unengaged, unsubscribed or bounced contacts count toward the allowance you are billed on is not published anywhere we could find, and it is the question we would put in writing first.

## The exit

This is where Keap separates itself from every other product in this category, and we want to be precise rather than dramatic about it.

Keap's own billing policy states that account cancellations must be initiated verbally through a Keap representative, that a request must arrive at least ten days before the billing date or it takes effect only at the end of the next cycle, and that emailed cancellation requests are not acceptable^[4]. There is no self-service cancellation of any kind.

The governing agreement adds the rest. Keap's terms-of-service address returns a redirect to the acquirer's terms, so the document a Keap subscriber signs is the Thryv Terms and Conditions, which we read in full^[5][6]. It gives a full refund only within three days of purchase and states that otherwise no refunds will be provided. It requires thirty days' notice before the end of the current period, which is not the ten days Keap's own pages ask for. And it makes termination effective on the expiry of your next subscription period after notice is received, which means correctly served notice ends the contract one full period later than most people would expect.

The annual contract also carries a $299 early termination fee, which the catalogue lists as its own line item with a billing trigger of contract termination^[2][3]. It is the only item in the catalogue whose billing event is the end of the relationship. It is also, by coincidence worth noticing, exactly one month of the entry plan.

Against all of that, one point in the vendor's favour, because it belongs here. A free self-service export of contacts exists and is documented, which is more than several competitors offer^[11]. Two caveats come with it. Notes and tags are not included, so the segmentation your account is built on does not leave with the contacts. And an account that cannot send marketing or transactional email may also be prevented from receiving its export download link, which puts the accounts most likely to be leaving in the worst position to collect their data.

◆Our scorecard

2.7/5

AI 2.4/5

Thin, and placed where it costs the vendor nothing. Two assistants are published: an AI Content Assistant that composes emails and an AI Automation Assistant that generates a campaign from a template. Both are writing aids. Nothing in the published feature set scores a lead with a model, forecasts a pipeline, summarises a conversation or recommends a next action, and the vendor's own lead scoring is described as rule-based prioritisation rather than as a prediction. The one genuinely useful thing about the placement is that it is uniform: because every plan carries the whole product, a buyer on the entry plan gets the same AI as a buyer paying twice as much. That is worth half a point here, and it is the only axis where the flat feature model helps rather than hurts.

Value for money 1.9/5

The lowest value mark we have given in this category, and the arithmetic is not close. The cheapest published way into Keap is $299 a month for 1,500 contacts and two user licences, plus an implementation package the vendor itself describes as required, which the catalogue prices between $1,000 and $3,000. A first year at the entry plan is therefore $4,588 rather than the $3,588 the monthly figure implies, and the number a buyer compares against other CRMs is 78% of the real one. Two things pull the mark up from where that would put it. The plans are not crippled: every feature Keap sells is on every plan, so nobody pays more to unlock functionality, which is rarer in this category than it should be. And because the price is per account rather than per seat, the comparison inverts with team size — two people on Keap pay far more than two seats elsewhere, while fifteen people on the entry plan plus thirteen licences land below what a mid-tier per-seat CRM would charge. What keeps the mark under two is that the included contacts are the most expensive contacts in the vendor's own catalogue: just under twenty cents each inside the entry plan against three and a half cents bought as an add-on. A buyer following the pricing page's slider upward is paying the worst rate available.

Capability & depth 3.8/5

The strongest axis, and the reason this product has survived twenty years. What Keap sells is not a CRM with marketing bolted on; it is a CRM, a marketing automation engine, an email and SMS platform, a landing page and form builder, a quoting and invoicing system, a checkout with order bumps and promotion codes, a native payment processor and an appointment scheduler, in one account. Measured against the other six CRMs in this category, that breadth is first or second on every comparison we can make from published feature lists. Two things hold it below four. The depth inside each area is shallower than a specialist's — the pipeline is a pipeline, not a forecasting system — and the automation library is metered: five, ten or fifteen prebuilt automations depending on plan, which is the only capability figure that moves with price in the entire product. A buyer who needs one deep thing should buy that thing; a buyer who needs eight adequate things in one place has few real alternatives at any price.

Ease of use 2.5/5

The measurement and the vendor's own behaviour point the same way, from opposite directions. In 98 English-language reviews read in full, complexity and interface together account for under 5% of the low scores — so the software is not the thing people are angry about, and the mark would be higher if reviews were the only evidence. But the vendor requires an implementation package and explains why in its own FAQ, which is as clear an admission as a vendor can make that the product is not self-service. Two published rules make that concrete: adding contacts and adding user licences both have the same documented answer, which is to reach out to a Customer Success Manager or support. A product where you cannot raise your own limits from inside the product has an ease-of-use ceiling that no interface can lift, and the three paid service ladders above the subscription — one-off implementation, monthly service packages to $3,297, and hourly data services — exist because enough customers need them.

Integrations 2.5/5

Scored on a count, not on an impression. The vendor's own marketplace, filtered with its own tabs, returns 88 applications against 255 experts: nearly three consultants for every integration, which is the shape of ecosystem a product grows when configuring it is the hard part. Eighty-eight native applications is at the low end of this category by a wide margin. What lifts the mark to the middle is the programmable surface, which is genuinely good and genuinely documented: 1,500 calls a minute and 150,000 a day per key, a published spike allowance, a documented 429 with retry guidance, plus native two-way mail and calendar sync with Gmail and Outlook and an embedded payment processor. Two things keep it from going higher. There is no published per-plan API entitlement anywhere, yet the catalogue sells 500,000 calls a month for $200 — a top-up to a quantity the vendor never states. And every deep link into the integration documentation now redirects to the acquirer's knowledge base front page rather than to the article, so the surface is easier to call than to look up.

User sentiment 2.1/5

A three-point spread, and for once the spread is explained rather than averaged. Trustpilot displays 1.1 from 498 reviews while that platform's own distribution table shows 62% five-star and 20% one-star, which averages 3.9: the gap is method, because the score weights recent reviews and the page itself discloses that the company has not recently invited customers. One Gartner Digital Markets pool, served identically by Capterra, Software Advice and GetApp, gives 4.1 from 1,299. TrustRadius gives 6.2 out of 10, which is 3.1. So the honest reading is not a number but a direction: older solicited reviews are positive, recent unsolicited ones are not. What the reviews are about matters more than where they land, and that was measured at both ends. Cancellation appears in 57% of the one and two-star reviews and in none of the four and five-star ones. Contracts, 41% against none. Billing, 39% against 5%. The mark is low because the negative mass is real, recent and one-directional, and it is not lower because almost none of it is about the software.

Docs & learning resources 2.3/5

The articles are good and you cannot get to them. On substance the knowledge base is better than most in this category: it publishes the unengaged-contact thresholds, the exact contact overage behaviour in both directions, the email ratio that forms the real sending ceiling, and what a contact export omits — the kind of operational detail several rivals simply leave out. The developer documentation is equally specific on rate limits and error handling. Against that, the knowledge base has been moved onto the acquirer's domain and every help.keap.com address now redirects to the new site's front page instead of the matching article, so seven years of links from the vendor's own pages, its community and its partners all land in the wrong place. Reading these articles at all required going around a filter that refuses ordinary requests. Documentation that exists but cannot be reached from the product that needs it is worth about half of what it should be, and the plan documentation proper is worse: the pricing page names no plan and prices none of the three it sells.

Support & reliability 3.3/5

The highest support mark we have given in this category, which surprised us and then held up. What is published is strong: United States phone support Monday to Friday from 6 a.m. to 6 p.m. Mountain time, chat around the clock inside the application, and a named Customer Success Manager on every plan including the entry one. What is measured agrees. Support is the dominant theme of the positive reviews, in 63% of the four and five-star reviews against 42% of the one and two-star ones, and the positive reviews are specific about it in a way that is hard to manufacture: they name a person and describe a problem that got fixed. Compare Freshsales, where support scored 2.2 and the ratio ran the other way. Two deductions keep this off four. The success manager a buyer is told to depend on — for adding users, adding contacts and cancelling — carries no obligation, no response time and no service level anywhere in the governing agreement; it is a feature bullet. And the same telephone that support answers well is the only channel through which an account can be closed, because the vendor's own billing policy states that emailed cancellation requests are not acceptable. Support that is excellent at helping and is also the only exit door is not the same thing as support that is simply excellent.

Weighted across the axes above (weights reflect what actually drives the decision in this category). How we score →

## Why the review scores disagree by three points

Trustpilot shows 1.1 from 498 reviews. Capterra shows 4.1 from 1,299. Both are real, and neither is the answer.

We read Trustpilot's own distribution table on the page: 62% of those 498 reviews are five-star and 20% are one-star, which averages 3.9^[17]. The displayed score is 1.1. The gap is method, not opinion. The score weights recent reviews, and the page itself discloses that the company has not recently invited customers, so a large mass of older solicited five-star reviews no longer carries the score while recent unsolicited one-star reviews do. Capterra's 4.1 is one review pool served identically by Software Advice and GetApp, with the same count and the same star split on all three, so it is one signal and not three^[18]. TrustRadius sits between them at 6.2 out of ten.

What the reviews are about settles it better than where they land. We read 100 of them in full and cross-tabulated themes against star ratings.

Theme	Share of 1-2 star reviews	Share of 4-5 star reviews
Cancellation	57%	0%
Contracts	41%	0%
Billing	39%	5%
Price	41%	16%
Support	42%	63%
Interface and complexity	under 5%	under 5%

Read the first row twice. Cancellation appears in more than half of the unhappy reviews and in not one of the happy ones. Support runs the other way: it is the dominant theme of the positive reviews, and those reviews are specific in a way that is hard to fake, naming a person and describing a problem that got fixed.

So the low scores are not a verdict on the software. They are a verdict on the exit, and they line up exactly with what the vendor publishes about how the exit works.

## The acquisition is visible in the plumbing

Keap belongs to Thryv, and four traces sit in the vendor's own properties. The terms-of-service URL redirects to the acquirer's domain^[5]. The help centre redirects to the acquirer's knowledge base, and not to the matching article but to its front page, so every deep link to Keap documentation is broken, including Keap's own. The support FAQ offers the acquirer's training academy^[2]. The catalogue carries the acquirer's cross-sell and implementation items^[3]. The vendor's acquisition FAQ puts continuity carefully: for now, all the Keap products and services you know and love will remain the same^[15].

For now is doing a lot of work in that sentence, and we would read it as written.

Two details a careful buyer will meet

There is a fourth plan in the catalogue. It sits at $299 a month and $2,988 a year, carries the same 1,500 contacts and two users as the entry plan, and has no prebuilt automation line at all^[3]. It is not offered on the pricing page and we found no vendor page explaining who buys it, so we record it as an observation of the catalogue rather than as something you can choose. If a quote arrives at the entry price with no automations attached, that is probably what you are looking at.

The other detail concerns long-standing accounts. The catalogue carries grandfathered items at no cost: free additional users, free contacts, legacy users and legacy contacts^[3]. Those terms are not available to a new buyer. Anyone who has been on this product since it was called Infusionsoft is therefore sitting on an arrangement a migration to a current plan would end, and the real cost of that move is not the headline difference between two prices.

## Where Keap actually fits

A business with an established customer list, a transactional model it wants automated end to end, and the willingness to be configured. If you need a CRM, an email engine, a checkout and an invoice in one place and you do not want to integrate four products, there are very few real alternatives at any price, and that is why this product has lasted twenty years.

A business that wants to try software, grow into it and change its mind later should look somewhere else. The entry cost is $299 a month plus a required package, the capacity cannot be raised or lowered without a phone call, and the way out is a verbal request with a fee attached.

## What we would put in writing before signing

The marketplace tells you something about the commitment. Counted with the vendor's own filters, it carries 88 applications against 255 experts^[16]. That is nearly three consultants for every integration, and it is the shape of ecosystem a product grows when configuring it is the hard part.

Our shortlist of questions, in order: whether unsubscribed, bounced and unengaged contacts count toward the billed allowance; what happens to contacts, automations and history on a downgrade, which the vendor's own upgrade article explicitly declines to cover; which notice period actually governs, the ten days on Keap's pages or the thirty in the contract; what the ninety-day money-back guarantee named on the legal index actually promises, given that the terms that index links allow a refund only within three days^[8]; and what the per-plan API entitlement is, because the catalogue sells a top-up of five hundred thousand calls a month for two hundred dollars to a base quantity that is published nowhere^[13].

Get those in writing. The product is good enough to be worth the paperwork.