Choosing an e-commerce agency is one of those decisions whose cost only becomes visible after the fact — usually as a fall in sales, a burnt budget and the need to do the same thing a second time. For a mid-market store (turnover of a dozen to several dozen million złoty a year) a bad decision can cost well over half a million złoty in the first year.
It is not one big invoice, but the sum of three things. Below are the five mistakes that most often lead to that scenario — and for each one a specific question with which to spot the danger before you sign the contract rather than after. Where does that order of magnitude come from? It accumulates from three components:
Lost traffic and sales. The industry benchmark for the drop in organic traffic after a badly executed migration is 15–30%. At mid-market turnover, where a significant share of sales comes from Google, that is hundreds of thousands of złoty a year.
A burnt implementation budget. A project that has to be rewritten after a year is not an investment — it is a down payment on a second project just like it.
The cost of the repair. Recovering Google rankings, fixing the architecture and moving to a new vendor usually cost more than doing it properly the first time.
You will recognise each of these five mistakes with one question asked before signing the contract. The worst thing for a mid-market store is to hear "that was not in the scope" only after the fact.
Mistake 1: Choosing on the lowest price
The cheapest offer almost always looks best in a spreadsheet — until it turns out what was left out of it. The worst investment in e-commerce is a store that has to be rebuilt after a year because the chosen architecture could not withstand the growth of the business. You then pay twice: once for a cheap launch and once for fixing it.
Price in itself is not the problem. The problem is a price without the context of what it covers and how long the solution is designed to last.
How to spot it before you sign: ask directly about the lifespan of the architecture — "how many years are you designing this store for?". A good answer is 5–7 years, not "for now". Also ask for the offer to be broken down into what is included in the price and what is outside it. If the second column does not exist, the extra costs will appear anyway — only after signing.
Cost of the mistake: rewriting the store within 12–18 months, that is the same implementation budget a second time.
Mistake 2: A migration without an SEO process
If you are planning to change platform, this is the mistake that costs the most and the fastest. Most agencies run a migration technically and treat SEO as a repair layer after launch — "we will do the redirects at the end". The result is typically a 15–30% drop in organic traffic, part of which never comes back.
The opposite approach — SEO-first, which we apply in every store migration — starts every sprint with an SEO decision and only then a technical one, with one-to-one URL mapping validated by an SEO specialist, not only by a developer. In our projects this keeps the traffic drop below 3%.
How to spot it before you sign: ask when work on SEO starts in the schedule — in the first sprint or after go-live? And whether a 1:1 redirect map is created for all indexed addresses, validated by an SEO specialist rather than merely generated automatically. "We will deal with that after launch" is a red flag.
Cost of the mistake: losing 15–30% of organic traffic is, at mid-market turnover, most often the single largest item on this list.
Mistake 3: Vendor lock-in and no ownership of the code and data
Some agencies build their model on the fact that you cannot leave them: code under their licence, a repository you have no access to, customer and order data "somewhere in the system", and reports in the form of decorative PDFs instead of real insight. As long as everything works, the problem is invisible. It becomes visible on the day you want to change vendor.
How to spot it before you sign: ask three questions. Who owns the code and the repository once the project is finished? Do I get a full export of store and customer data on request? What does my ongoing insight into the work look like — real visibility of hours used and task status, or a monthly PDF "decorative for marketing"? A lack of a clear answer to any of them is a warning sign.
Cost of the mistake: being trapped with one vendor means you pay their rates with no real alternative, and leaving — if it is possible at all — requires another migration.
Mistake 4: Selling hours instead of a business result
The "you buy a package of hours" model looks fair, because you pay for work performed. The problem is that it shifts all the risk onto you: the agency earns the same regardless of whether the hours translated into sales. Nobody on the other side has any reason to work faster or more effectively.
A sign of a mature agency is a willingness to tie part of the guarantees to the result — an SLA on the traffic drop after a migration, contractual penalties for delays, concrete commitments instead of promises. That is skin in the game: if someone promises something, they should bear the consequences of not delivering.
How to spot it before you sign: ask what the agency commits to contractually beyond "working the hours". Is it prepared to write in a guarantee for the outcome of the migration or a penalty for delay? If the only unit of billing is the man-hour, you are buying activity, not a result.
Cost of the mistake: the budget spent in full, the result arbitrary. It is also worth knowing that a store doing PLN 10m+ a year usually outgrows billing by the hour and needs a model geared towards growth.
Mistake 5: An agency with no experience in your platform or in mid-market migrations
A generalist who "does everything" learns your platform at your expense. Mid-market e-commerce has its own specifics — a B2B store on Sylius has different requirements from a D2C one on Shopify Plus, and a legacy migration from PHP 5.x to Symfony different ones again. Experience in exactly that path determines whether the architecture is right from day one.
Paradoxically, a good sign is an agency that can say when it is not for you — and point you to a better vendor outside its area. Those who know the limits of their specialisation deliver what they promise within it; we describe that philosophy in more detail on the about us page.
How to spot it before you sign: ask for case studies from your exact path — the same platform, a similar catalogue scale, the same type of migration. Also ask what kind of projects the agency does not take on. The answer "we do everything" is itself an answer.
Cost of the mistake: a wrong architectural decision at the start comes back as rewriting the store — that is, mistake number 1 from this list, only a year later.
Cheat sheet: five questions before signing
If you are to remember one thing from this whole text, let it be this table — five questions and five red flags that separate a good agency from a costly mistake.
MistakeControl questionRed flag in the answerChoosing on priceHow many years are you designing this store for? What is outside the price?"For now"; no breakdown of scopeMigration without SEOWhen does SEO start and is a 1:1 redirect map created?"We will deal with that after launch"Vendor lock-inWho owns the code and the data? What insight do I have into the work?No data export; reporting only as a PDFSelling hoursWhat do you commit to contractually beyond hours?The only unit of billing is the man-hourNo specialisationShow me a case from my platform and scale. What do you not take on?"We do everything"
Summary
A bad choice of agency rarely shows up immediately — it accumulates over the first year in lost traffic, a burnt budget and the cost of repair, which together can exceed half a million złoty. The good news is that you will recognise all five mistakes before signing the contract, by asking a few specific questions and listening for whether the answers are unambiguous. An agency that is not afraid to say what it does not do and what it takes contractual responsibility for is usually a safer choice than the one with the lowest offer.
If you are facing such a decision — especially with a migration — start with an independent store audit, which will give you hard data for a conversation with any vendor. We are also happy simply to talk about your case and say honestly whether we are the right partner for it.