How a Window Service Company Overcame a Crisis and Increased Profits by 230%: Case Study of Atlant Digital and AX Services
Alex Hristich
??CEO of Atlant.Digital agency ??8 years of experience in performance marketing
What to do if the cost per customer in PPC advertising has skyrocketed, and the landing page has started converting poorly? Increasing the budget is not an option — it’s not unlimited. Customer acquisition costs exceed profits. Everything is going downhill. What’s the solution? Gather all your strength and prepare for a gradual analysis and optimization.
In this case, you’ll learn how a small company in Kyiv managed to bounce back from losses, exceed $6900/month in revenue, and take a leading position in the window services market in their region.
About the Project ??
Two window repair technicians from Kiev named Oleksandr and Maksym made the decision to launch a small company. A year prior, all was good with the world: their one-page website was generating respectable traffic from Google Adwords ads. Each day, they received two to three orders, so they had plenty work. But precisely three months ago, everything began to go south quickly: orders stopped pouring in, advertising expenses skyrocketed, and now a click costs 20 UAH instead of 8 UAH. The landing page was no longer functional. The objective was to fulfill the same amount of orders within the same budget, if not more. The objective was to raise the company’s monthly revenue to $2,000 at most.
Our Strategy ??
? What We Did:
? Reconfiguring Advertising Campaigns: Since the guys were not experts in setting up PPC ads, their account was a complete mess:
There was no point in fixing the mess — it needed a complete overhaul.
? Steps for Setting Up New Advertising Campaigns: We broke down all ads by service type:
We structured the ads by service type to:
Analyzing Competitors:
When your ad performance and landing page efficiency suddenly plummet, two common hypotheses emerge:
? Competitor analysis is the first step when developing a marketing strategy for a business. It helps to understand:
We compiled a list of all competitors and analyzed their ads using Serpstat and SimilarWeb tools. We also reviewed their landing pages and consolidated the data into a single table.
? What We Learned:
Our conclusion: If we just do everything well, it will pay off.
Keyword Research:
I won’t dwell too much on this, but here are the key steps:
Ad Creation:
In our team, we follow the method of 1 keyword = 1 ad group. This is the best approach I’ve tried, although for large-scale keyword lists, you may need to group multiple keywords in one ad group.
For each ad group, we create 3 types of ads:
This method allows for quick testing of different ad texts to identify the highest-converting ones.
? Setting Up Advertising Campaigns:
Apart from all the basic settings “as Anton Petrenko recommends”:
? Creating Landing Pages
Problems with the previous landing page:
Niche Analysis/Client Brief
Before creating the landing page, it’s crucial to understand the target audience and the specific functions the page needs to serve.
After a briefing and discussion with the client, we identified the main objections and needs of their customers:
1. Implementing Benefits and Addressing Objections on the Page
2. Creating Separate Pages for Each Service
In PPC advertising, an enticing ad motivates the click. But if someone clicks on “pizza delivery in Kyiv with a 15% discount” and lands on a homepage for sushi delivery, even the best ad won’t succeed.
Ensuring the landing page content matches the user’s search query boosts conversion rates. This is the principle of relevance. It’s simple psychology: we respond more favorably to something that is directly related to our needs.
Therefore, we created separate pages for each service, and using dynamic content, we adjusted the headlines and text on the first screen based on the keyword that brought the user to the page.
领英推荐
? Creating Landing Pages
Problems with the previous landing page:
Niche Analysis/Client Brief
Before creating the landing page, it’s crucial to understand the target audience and the specific functions the page needs to serve.
After a briefing and discussion with the client, we identified the main objections and needs of their customers:
1. Implementing Benefits and Addressing Objections on the Page
2. Creating Separate Pages for Each Service
In PPC advertising, an enticing ad motivates the click. But if someone clicks on “pizza delivery in Kyiv with a 15% discount” and lands on a homepage for sushi delivery, even the best ad won’t succeed.
Ensuring the landing page content matches the user’s search query boosts conversion rates. This is the principle of relevance. It’s simple psychology: we respond more favorably to something that is directly related to our needs.
Therefore, we created separate pages for each service, and using dynamic content, we adjusted the headlines and text on the first screen based on the keyword that brought the user to the page.
Exact alignment between the search query and the content on the landing page increases its conversion rate. This is the principle of relevance. In reality, it’s simple psychology: we are more inclined to respond positively to something that feels personally connected to us.
Therefore, not only did we create separate pages for each service, but we also used dynamic content to adjust the headlines and text on the first screen based on the keyword that led the user to the page.
3. Sequential A/B Testing of Landing Pages
Achieving high conversion rates instantly was not feasible, so for each service, we conducted at least two A/B tests. For the most visited site, we tested as many as six variations.
What We Tested:
We achieved strong results:
? Optimization of Advertising Campaigns
Over the course of a month, our goal was to reach 15 inquiries per day. Thanks to a patient client and our team’s diligent work on optimization, we finally succeeded. But let’s skip the pleasantries and focus on what we actually did.
What we used for optimization:
The most basic optimization steps, what we call “account hygiene,” include turning off ineffective keywords and updating the negative keyword list.
1. Rewriting Ads and Disabling Ineffective Ones
Since we had three ad variants with different text, it was easy to identify which ads were effective and which were not.
By keeping only the best-performing ads, we reduced the cost per lead by 34%.
2. Running Experiments with Different Ad Settings
We primarily tested various bidding strategies. Some strategies worked exceptionally well for certain campaigns, while others didn’t. This made experiments an essential tool.
For those unfamiliar, an experiment in Google Ads is a built-in tool that allows you to conduct A/B tests.
3. Manually Adjusting Bids for Each Keyword
A huge thanks to Oleksandr Holubenko, who spent hours each day fine-tuning bids for maximum accuracy. Some might argue that this is a waste of time and that automated bidding tools could handle it, but unfortunately, automated tools aren’t yet capable of understanding customer psychology with the same precision.
4. Implementing Remarketing
Unfortunately, our remarketing efforts in Google Ads didn’t work out as planned. The cost per lead was three times higher than our KPI.
Firstly, people make decisions very quickly, so if they see your ad later in the day, it may already be irrelevant.
Secondly, remarketing should be used to simplify the process for users, either by offering an easier step (like a lead magnet) or discounts and promotions. Since we had neither, we decided to turn it off.
5. Automated Rules and Bid Adjustments
From our time-of-day report, we identified the most competitive periods where our CTR and positions were falling, and we increased bids by 15–20%.
Campaign optimization became the accelerator for the results we achieved one month after launching the ads. Patience and persistence are key when striving to improve ad performance to its highest potential.
Outcomes ??
Following three months of collaboration with the client, we achieved a $7,000 monthly turnover and a 330% return on investment. There are currently two supervisors, five artisans, and a CRM system in operation inside the team.