9 out of 10 requests in our B2B niche turned out to be fake. How we almost lost our SEO budget trying to sell software to couriers

We are developing adelivo.ru - a system in which the delivery service sees orders, couriers, routes and finances in one window. Before making a major website update and pouring budget into SEO, we decided to look at the niche through the eyes of search: what people actually type when they need a system like this.

The answer turned out to be extremely inconvenient. The largest flow of requests in our niche is not a business looking for software for a courier service. These are couriers looking for work.

We'll tell you how we revised the strategy in time, why cool texts can work for zero, and where the money actually lies in B2B delivery.

Numbers that made it necessary to redo the plan

If you just take the basic keywords of our niche and look at their frequency, the picture seems attractive. Here is the top of the demand:

And here’s what real B2B demandlooks like - that is, people who already have orders, couriers and the need for automation:

| courier routing | 45 |Break in order.

If you blindly chase frequency, you can bring several thousand visits to the site in a couple of months and receive exactly zero B2B applications. People will come to you who need a weekend shift, not a SaaS platform for the company's money.

We didn’t get burned by this only because we did the math before we ordered the texts. Usually they do the opposite: they write articles, traffic grows, there are no sales, and the classic conflict begins: “bad SEO - bad product.”

One word that cost us four pages

Then things got even more interesting. We decided to check the wording that we used to call our landing pages. Spoiler: we were in for a surprise.

Zero. Not “little” - but absolute, crystal clear zero. Wordstat replied: “no matching queries were found.”

We didn't believe it at first. We decided that the service was buggy or the page didn’t load. The test turned out to be simple: they typed the obviously meaningless “ytsukengshschz fyvapr” and compared the answer. He was exactly the same. This means our zero is honest.

Now let's look at that same keyword:

“By couriers” – zero. “Delivery” - 793.

The difference is only one word, but it was on the word “couriers” that we built the best page of the site: almost ten thousand characters of good, verified text, aimed at absolute emptiness.

• Honest disclaimer: without it, the number 793 will be a lie. A significant share of these impressions are about marketplaces (“ozone delivery management system” - 332). These are people who are looking for a PV account, not software for their service. There is less than half of pure targeted B2B demand here. But even so it is not comparable to zero.*

Where in the niche the requests actually lie

The most “monetary” find turned out to be not in general thematic queries at all, but in branded ones - that is, in the names of third-party systems with which we can integrate.

For comparison, the general target query “crm for delivery” is searched only 69 times. The difference is 27 times.And most importantly: the person who types “retailcrm” is an already existing business. He has orders, processes and a budget for software. He's definitely not looking for a part-time job as a courier. He is looking for how to connect his current CRM system with logistics and delivery - that is, exactly what our platform solves.

The conclusion was unexpected for us: high-quality B2B applications will most quickly come not from the rewritten main page, but from the integration pages. RetailCRM, Bitrix24, Console.Pro, Yandex Maps, Telegram. We already had these pages, we just needed to smartly tailor them to brand demand.

What we did with it: 3 steps

Having realized the real picture, we turned the order of work 180 degrees.

1.We rewrote H1, Title and Description for all 22 pages of the site.We started with those that targeted dead wording. This is not a cosmetic change: four pages of our website previously looked at zero demand. 2.Changed priorities.The first in line for global content revision was the RetailCRM integration page, not the main one. 3.We divided the analytics.We left the page for couriers-applicants and don’t even touch it - our clients also need couriers, the flow from it is useful. But we now consider its traffic as a separate line and do not include it in the KPI of the main B2B promotion. Otherwise, the traffic report will be beautiful, but the number of applications for the software itself will not increase.

This third point seems like an obvious detail, but it is where half of the marketing reports in niches with two-sided demand break down: applicant/performer traffic is mixed with customer traffic. The overall figure on the graph is growing, everyone is happy, but the business does not understand why, with such growth, the sales department’s phone number is silent.

Why this applies not only to delivery (results)

Two-sided demand (marketplace model) exists wherever a platform or business brings together two sides: cleaning, cargo transportation, tutors, repairs, medicine, taxi. And everywherethe side is looking for performers more often and more actively than the side of customers. There are simply more of them physically, and they are looking for work/part-time work every day.Three things to do before blowing your content and SEO budget:1.Split demand by intent, not topic.One request, one party. “App for couriers” and “delivery management system” lie in the same semantic niche, but lead to completely different people with different goals. 2.Check the wording of your pages right now.We found four landing pages aimed at zero demand. Zero in Wordstat is not just an empty cell in a table. This is the money you pay a copywriter for text that no one will ever read. 3.Look at brand requests of related systems. In integrations there is an audience that already pays for software. And there is an order of magnitude more of it than in direct commercial requests for your narrow topic.

• At ADelivo we make sure that delivery services work without chaos. Orders, couriers, AI routing and payments in one window. Just to understand the load: as of March 8, in one of our implementations, 700 orders per day were smoothly processed through the system with 100 couriers on line.*