The difference is what you own at the end
Strip away the arguments about cost and it comes down to one distinction.
When you pay a portal, you are renting access to demand that the portal owns. The enquiry is theirs and they lend it to you. Stop paying on a Monday and by Tuesday you have nothing — not a page, not a ranking, not a review you can take with you, not a customer list. Six years of spend leaves exactly the same position as six weeks.
When you build your own, you are buying an asset. Slowly, and it is genuinely slower than anybody selling it to you will admit. But a page that ranks keeps ranking without further payment, a Google review stays on your profile, and a customer who found you directly came to you rather than to a marketplace that happened to list you.
That is not an argument for never using a portal. It is an argument for knowing which one you are doing, because a lot of trades believe they are building something while they are renting.
The shared-lead problem
Portal models differ — some are directories you subscribe to, some sell individual leads, some take a commission on the job — and the fee structures change often enough that quoting a figure here would mislead you within a year. Check their current pricing yourself.
What is structural, and worth understanding whichever model you are on, is what happens to a lead that is sold more than once.
Where a lead goes to several trades at the same time, you are not competing on whether you are any good. You are competing on who rings back first and who comes in cheapest, because the customer now has three quotes for an identical job and nothing to distinguish them except price and speed. That is a race you can win, but winning it costs margin on every single job, for ever.
Contrast the enquiry that comes through your own site. That customer has read a page about the exact job they want, seen photographs of work you did two streets away, and formed a view about you before making contact. They are not comparing three names from a list. They are often not getting another quote at all.
This is the part that gets missed when trades compare cost per lead. Two leads at the same price are not the same lead if one arrives pre-sold and the other arrives with two competitors attached.
Reviews you cannot take with you
This one is quiet and it compounds badly over years.
Reviews left on a portal live on the portal. They are a genuine asset while you are paying, and they are worth nothing to you the day you leave. A trade with two hundred reviews on a platform and eleven on their Google Business Profile has spent years building somebody else's business.
Google reviews behave the opposite way. They attach to your business, they feed the map pack that sits above every organic result, and they are visible to somebody searching your name to check you out — which, in trades, is almost everybody before they let you into the house.
The practical move is not to stop asking on the portal. It is to ask for both, every time, at the moment the customer is happiest. That is when they have just seen the finished work, not three weeks later in an email. Trades that ask at the right moment get reviews; trades that email later mostly do not, and conclude that their customers are not the reviewing type.
There is a separate guide on this site about getting Google reviews properly, including what you may and may not do when asking.
The comparison people never actually run
Almost every trade compares these on cost per lead, which is the wrong denominator. The number that matters is cost per job won, and it is easy to work out once you decide to.
For the portal: what you paid in a quarter, divided by the jobs that came from it. Include the leads that went nowhere, the duplicates, and the ones that were never real — they were all paid for.
For your own channel: what you spent on the site and the hours you put into the profile over the same period, divided by the jobs that came from it. Spread the build cost over a couple of years rather than charging it all to one quarter, because the site does not stop working in month three.
Two things usually fall out of that for the first time. The portal number is higher than people assume, because the unconverted leads were invisible. And the owned number starts high and drops every quarter, which is exactly what you would expect from an asset and exactly what a rented channel never does.
You cannot run this calculation at all unless you know which enquiries came from where, which brings us to the thing that has to come first.
None of this is answerable without tracking the phone
Trade enquiries arrive by phone. If a tap on your number does not fire a tracked event, every figure you have describes only the people who filled in a form, and in most trades that is the minority.
It is a few lines of code on the phone link. It takes an hour. Almost no trade site has it, and without it the portal-versus-site argument is unresolvable, because you genuinely do not know where your work came from — you only know where you remember it coming from, which is a different and much less reliable thing.
Add to that a question on every call: "how did you find us?" Written down, every time, for three months. It is crude and it is better than nothing by an enormous margin.
Once both are running, this stops being an argument and becomes arithmetic.
Using the portal as a bridge
Here is the honest recommendation, which is not the one a web studio is supposed to give.
If the diary has gaps next month, pay the portal. Cash flow is not a strategy problem, it is a this-month problem, and a website will not fix it by Friday. Anybody telling a trade with an empty diary to cancel their lead spend and wait six months for SEO is giving advice that works out well for the person giving it.
What matters is what else happens in those six months:
- The Google Business Profile completed properly — categories, service areas, true hours, real job photographs added regularly. Free, and for a trade it outperforms the first several months of website work.
- Every customer asked for a Google review at the moment they are happiest, including the ones who came from the portal. Those leads are paid for; the review is the only part of them you get to keep.
- Phone-tap tracking on the site, so the comparison becomes measurable.
- One proper page for the highest-margin service you want more of, then one for the town you do most of that work in. Not twenty — two, done well.
- The cost-per-job-won number for both channels, recalculated each quarter.
Then the decision makes itself. When your own channel is cheaper per job won and producing enough of them, you reduce the portal spend — gradually, not in one go. If it never gets there, you have at least found that out with numbers rather than by arguing about it in the van.
What to work out this month
- What did you pay every lead source last quarter, in total?
- How many jobs did each one actually produce? Cost per job won, not per lead.
- How many reviews do you have on platforms you pay for, against on your own Google Business Profile?
- If you cancelled every paid source tomorrow, how many enquiries would arrive next month?
- Does a tap on your phone number fire a tracked event?
- Are you asking portal customers for a Google review as well?
- Which single service page would you write first, if you wrote one?
Question four is the one worth sitting with. If the honest answer is "almost none", that is not an argument for cancelling anything — it is the measure of how much of your business you are currently renting.
