Every blog writing tool on the market produces competent English. That happened somewhere around 2024 and it means the thing most people judge these tools on — does the writing read well — no longer separates anything. They all pass. Then you publish forty posts and nothing happens.
We build one of these tools, so treat this as an interested party's opinion. It is still the advice we would give someone shopping, including in the places where it does not flatter us.
The short answer
Four things determine whether a blog tool produces pages that earn traffic: whether it researches real search demand before writing, whether it knows anything specific about your business, whether each post has a destination, and whether you can see afterwards what happened. Writing quality is a fifth thing, and it is the one you can safely assume.
Does it research before it writes, or just write?
This is the single biggest divide, and it is usually invisible on the pricing page.
A tool that only writes takes your topic and produces an article about it. If you chose a topic nobody searches for, you get a well-written page with no audience, and you will not find out for four months. A tool that researches first goes looking for what people actually type, then tells you which of those phrases are worth a page.
You can spot the difference in the interface. If the first thing it asks for is a title or a topic, it is a writing tool. If the first thing it asks for is your website or a page URL, it is doing something more interesting before the writing starts.
Neither is wrong — a writing tool is genuinely useful if you already know your topics from talking to customers all day. But you are then doing the hard half yourself, and you should not pay for research you are not getting.
Does it know anything about your business?
Ask a general chatbot to write about gutter cleaning and you get an article about gutter cleaning. It is accurate, it is readable, and it is identical to what your competitor two suburbs over would get from the same prompt. Nothing in it could only have come from you.
That sameness is the actual problem with cheap content, far more than quality. A page that a competitor could republish unchanged has no reason to outrank theirs.
So look for what the tool takes in: your prices, your service area, your process, the objections you hear on the phone. Some tools read your existing site and pull this out automatically. Others ask you. Both work. What does not work is a tool that never asks, because then the specifics simply are not there.
Does each post have somewhere to send people?
A post that ranks and leads nowhere is a brochure someone read in a waiting room. Pleasant, forgotten.
Every article should be pointed at a page where the reader can do something — book, enquire, see prices, read the service detail. That link should be in the body where the reader is already convinced, not stranded in a footer. If the tool has no concept of a target page, you will be adding those links by hand on every post forever, and by post twelve you will stop.
This is also the difference between a blog that feels like marketing and one that feels like a diary. The diary version is the one most small business blogs become.
Can you find out whether it worked?
Most tools hand you a document and end the relationship there. You are then reliant on Google Analytics to tell you whether any of it mattered, and Analytics is genuinely hard to read on low traffic — forty visits a month across ten posts looks like noise no matter how you chart it.
What you want to know is narrower and more useful: did anyone who read the post click through to the page it was pointing at? That single number tells you whether the article did its job. Ten readers and three click-throughs is a good post. Four hundred readers and no click-throughs means you wrote something interesting about the wrong thing.
What no tool can do for you
Worth being clear about the ceiling, because the marketing around these products tends not to be.
No tool knows your business better than you do, and none of them will catch it when a draft states a price, a timeframe or a regulation confidently and wrongly. That happens often enough — roughly one post in ten will contain a detail worth correcting — that a human read before publishing is not optional. It takes about ten minutes. If you will not do those ten minutes, buy fewer posts rather than skipping the read.
No tool will fix a website that loads slowly, has no clear service pages, or does not say where you are based. Content sits on top of those things and cannot compensate for them.
And no tool makes this fast. Search takes months regardless of who writes the words. Anything promising results in weeks is describing paid advertising, not blogging.
A twenty-minute test before you pay for anything
Most of these tools offer a free or cheap first post. Use it properly rather than skimming the output and thinking "yes, that reads fine".
- Give it your real website and a real page you want enquiries for. Not a test topic. The whole point is to see what it does with your actual business.
- Read only the first paragraph. Does it answer a question, or does it clear its throat for sixty words? The opening is where most generated content gives itself away.
- Hunt for one fact that could only be about you. Your suburb, your price range, your process, your seasonality. If there is nothing, the tool never asked and never looked.
- Check where the links go. Is there a link to the page you nominated, sitting somewhere a reader would naturally click?
- Search the topic yourself. Type the question into Google. If the first page is all national directories and companies with marketing departments, the tool picked a fight you cannot win — which tells you its research is weak.
A tool that passes those five is worth paying for. One that produces beautiful prose and fails three of them will cost you a year.
If you are publishing for clients
Web designers, hosting companies and agencies have a different problem: the tool has to work across thirty sites without thirty separate logins and thirty manual copy-pastes. The per-post price stops mattering and two other things start to.
The first is how content gets onto the site. If each post needs someone to log into WordPress, paste HTML and fix the formatting, that is fifteen minutes per post per client and it does not scale past about ten clients. An embed or an export that drops in cleanly is the difference between a service you can sell and a chore you resent.
The second is whether you can show the client anything at the end of the month. "We published four posts" is not a report. "These four posts sent eleven people to your quote page" is a renewal.
Where we sit in this
Blog Lab does the four things above: it reads your site, researches search phrases for a page you nominate, writes long-form posts linked back to that page, and tracks who clicked through. It is built for New Zealand and Australian service businesses and for the people who publish on their behalf. You can see the whole process before you spend anything.
What we are not is a replacement for a copywriter on the pages that matter most, and we would rather say so than have you find out on your homepage. We have written separately about when to hire a person instead, including the cases where the answer is clearly not us.
The honest summary is this: the writing is the easy part now, and any tool charging you mainly for the writing is charging you for the commodity. What is still hard is knowing what to write about, making it specific to one business, pointing it somewhere, and proving it did something. Judge the tools on that.