[{"data":1,"prerenderedAt":893},["ShallowReactive",2],{"navigation-en":3,"language-switcher-\u002Fen\u002Fblog":87,"blog-index-en":93},{"services":4,"industries":48,"caseStudies":69},[5,10,14,19,23,27,31,35,39,43],{"slug":6,"navLabel":7,"navGroup":8,"heroHeadline":9},"web-applications","Web applications","build","A web application built around your process",{"slug":11,"navLabel":12,"navGroup":8,"heroHeadline":13},"erp-systems","ERP systems","An ERP system built for your company, not the other way round",{"slug":15,"navLabel":16,"navGroup":17,"heroHeadline":18},"integrations-and-api","Integrations and APIs","improve","Your systems stop being islands",{"slug":20,"navLabel":21,"navGroup":17,"heroHeadline":22},"ai-implementation","AI implementation","AI deployed where it actually saves time",{"slug":24,"navLabel":25,"navGroup":17,"heroHeadline":26},"process-automation","Process automation","Work no one should be doing by hand",{"slug":28,"navLabel":29,"navGroup":8,"heroHeadline":30},"websites","Websites","A site that loads instantly and brings in inquiries",{"slug":32,"navLabel":33,"navGroup":8,"heroHeadline":34},"mobile-applications","Mobile applications","One codebase, both platforms",{"slug":36,"navLabel":37,"navGroup":17,"heroHeadline":38},"code-audit-and-project-takeover","Code audit and takeover","The project stalled. We'll take it over and tell you the truth about its condition",{"slug":40,"navLabel":41,"navGroup":17,"heroHeadline":42},"maintenance-and-support","Maintenance and support","A system needs someone watching it after launch",{"slug":44,"navLabel":45,"navGroup":46,"heroHeadline":47},"discovery-workshop","Discovery workshop","start","Before you decide to build, find out exactly what you're building",[49,53,57,61,65],{"slug":50,"navLabel":51,"heroHeadline":52},"manufacturing","Manufacturing","The works order reaches the floor as a printout from an email",{"slug":54,"navLabel":55,"heroHeadline":56},"automotive","Automotive","The customer exists in three places at once",{"slug":58,"navLabel":59,"heroHeadline":60},"industrial-automation","Industrial automation","You sell competence that doesn't photograph",{"slug":62,"navLabel":63,"heroHeadline":64},"e-commerce","E-commerce","The shop works. What sits behind it comes apart",{"slug":66,"navLabel":67,"heroHeadline":68},"logistics","Transport and logistics","The job arrives by email, the status comes back by phone",[70,77,82],{"slug":71,"client":72,"tag":73,"year":74,"summary":75,"image":76},"dss-by-caredo","Caredo","Product",2025,"We build the whole DSS product together with the Caredo team — from the CRM and planner to the software behind the key lockers and totems.","\u002Fimg\u002Fcase-studies\u002Fdss\u002Fkluczykomat\u002Fframe_0169.webp",{"slug":78,"client":72,"tag":79,"year":74,"summary":80,"image":81},"dss-by-caredo-website","Product website","We built the DSS product website end to end — from design to launch — on Caredo's visual identity.","\u002Fimg\u002Fscreenshots\u002Fdss-desktop.jpg",{"slug":83,"client":83,"tag":84,"year":74,"summary":85,"image":86},"2motion","Brand and website","A whole brand from scratch — logo, colours, business cards — a company website and blog posts that position 2motion in search.","\u002Fimg\u002Fscreenshots\u002F2motion-desktop.jpg",[88],{"code":89,"label":90,"name":91,"to":92},"pl","PL","Polish","\u002Fblog",[94,379,666],{"id":95,"title":96,"author":97,"body":98,"category":360,"date":361,"description":362,"draft":363,"editorialNote":364,"extension":365,"image":366,"meta":367,"navigation":368,"path":369,"rawbody":370,"seo":371,"sitemap":372,"stem":373,"tags":374,"translationKey":377,"updated":364,"__hash__":378},"blogEn\u002Fblog\u002Fdedykowany-erp-czy-gotowy-system.md","Dedicated ERP or off-the-shelf — how to choose","zespol",{"type":99,"value":100,"toc":344},"minimark",[101,110,113,118,123,126,130,133,137,140,144,147,151,154,248,255,259,278,281,285,292,295,298,302,308,314,320,324,327,334,341],[102,103,104,105,109],"p",{},"The answer is: ",[106,107,108],"strong",{},"off-the-shelf, in most cases",". Worth stating up front, since a piece written by a company that builds dedicated systems might be expected to argue the opposite.",[102,111,112],{},"An off-the-shelf ERP wins when your process resembles other companies' in the industry and the budget is limited. Licence plus rollout is then cheaper than building, and the risk is lower. A dedicated system makes sense in a narrower set of situations — and this piece is about recognising when you're in that set.",[114,115,117],"h2",{"id":116},"four-deciding-questions","Four deciding questions",[119,120,122],"h3",{"id":121},"_1-is-your-process-a-source-of-advantage","1. Is your process a source of advantage?",[102,124,125],{},"If a company makes money because it does something differently from its competitors, flattening that \"differently\" into a standard module costs more than the licence. If, on the other hand, the process is standard and the advantage lies elsewhere — relationships, price, location — an off-the-shelf system is the right choice.",[119,127,129],{"id":128},"_2-what-does-adapting-the-off-the-shelf-system-cost","2. What does adapting the off-the-shelf system cost?",[102,131,132],{},"This question settles it most often. Ask the vendor for a quote on the modifications you need. If the sum of licence, rollout and modifications approaches the cost of building your own system, the choice stops being obvious — because for the same money you get either an 80%-fit system, or a 100%-fit one.",[119,134,136],{"id":135},"_3-how-many-users-do-you-have","3. How many users do you have?",[102,138,139],{},"A per-user licence changes the maths over time. At twenty people the difference is negligible. At two hundred, over a five-year horizon, it can exceed the cost of building your own system.",[119,141,143],{"id":142},"_4-how-much-will-the-process-change","4. How much will the process change?",[102,145,146],{},"A company in a stable industry, doing the same thing for a decade, fits well into an off-the-shelf system. A company changing its business model every two years will keep fighting limitations that can't be worked around except through the vendor.",[114,148,150],{"id":149},"the-five-year-comparison","The five-year comparison",[102,152,153],{},"Comparing only the rollout cost is misleading. Below is an example for a manufacturing company with 120 users. The figures are illustrative — meant to show the cost structure, not a quote.",[155,156,157,173],"table",{},[158,159,160],"thead",{},[161,162,163,167,170],"tr",{},[164,165,166],"th",{},"Line item",[164,168,169],{},"Off-the-shelf ERP",[164,171,172],{},"Dedicated",[174,175,176,188,199,210,220,231],"tbody",{},[161,177,178,182,185],{},[179,180,181],"td",{},"Rollout \u002F build",[179,183,184],{},"PLN 250k",[179,186,187],{},"PLN 450k",[161,189,190,193,196],{},[179,191,192],{},"Licences over 5 years",[179,194,195],{},"PLN 430k",[179,197,198],{},"0",[161,200,201,204,207],{},[179,202,203],{},"Modifications and extensions",[179,205,206],{},"PLN 180k",[179,208,209],{},"PLN 150k",[161,211,212,214,217],{},[179,213,41],{},[179,215,216],{},"PLN 120k",[179,218,219],{},"PLN 220k",[161,221,222,225,228],{},[179,223,224],{},"Infrastructure",[179,226,227],{},"included in licence",[179,229,230],{},"PLN 90k",[161,232,233,238,243],{},[179,234,235],{},[106,236,237],{},"Total",[179,239,240],{},[106,241,242],{},"PLN 980k",[179,244,245],{},[106,246,247],{},"PLN 910k",[102,249,250,251,254],{},"The takeaway isn't \"dedicated is cheaper.\" It's: ",[106,252,253],{},"at this scale the costs converge, so the decision should be made on a different basis than price",". At thirty users the same table looks completely different, and the off-the-shelf system wins outright.",[114,256,258],{"id":257},"signals-that-an-off-the-shelf-system-has-stopped-fitting","Signals that an off-the-shelf system has stopped fitting",[260,261,262,266,269,272,275],"ul",{},[263,264,265],"li",{},"The team keeps the \"real\" data in a spreadsheet next to the system.",[263,267,268],{},"A key report is produced by exporting to Excel and stitching it together by hand.",[263,270,271],{},"Every process change requires a ticket to the vendor and waits for quarters.",[263,273,274],{},"You're paying for modules nobody ever turned on.",[263,276,277],{},"A new hire is learning less the system than the list of workarounds the team invented.",[102,279,280],{},"One of these signals is normal friction. Three or more mean the cost of the mismatch has exceeded the cost of change.",[114,282,284],{"id":283},"a-third-path-thats-rarely-discussed","A third path that's rarely discussed",[102,286,287,288,291],{},"The choice isn't binary. A common and underrated approach is to ",[106,289,290],{},"keep an off-the-shelf ERP as the finance-and-accounting core and build your own modules where the company is genuinely atypical",".",[102,293,294],{},"Accounting, HR and payroll are regulated areas where an in-house build is an expense with no return — the rules change every year and someone has to keep up. A production module, or handling a specific sales channel, is exactly where a custom fit pays off.",[102,296,297],{},"The condition is an ERP with a decent API. Without one, this path is closed — and that's an important criterion in choosing an off-the-shelf system that almost nobody asks about before signing.",[114,299,301],{"id":300},"what-not-to-weigh-in-the-decision","What not to weigh in the decision",[102,303,304,307],{},[106,305,306],{},"\"Everyone in the industry has X.\""," Your competitors made that decision at a different scale, in a different year, with a different process.",[102,309,310,313],{},[106,311,312],{},"The sales demo."," Every system looks good in a demo the vendor prepared. Ask for access to a test environment and run your own real case through it — ideally the most unusual one you have.",[102,315,316,319],{},[106,317,318],{},"The promise that \"it'll get configured.\""," Ask who will configure it, how long it'll take, and what a consultant's day costs. The answers are sometimes surprising.",[114,321,323],{"id":322},"how-to-make-this-decision-properly","How to make this decision properly",[102,325,326],{},"Do two things before you choose.",[102,328,329,330,333],{},"First: ",[106,331,332],{},"write down your process as it actually is",", workarounds included. Those workarounds are information about what the previous tool was missing — and the best material for testing candidates.",[102,335,336,337,340],{},"Second: ",[106,338,339],{},"ask both types of vendor to quote the same, written-down scope",". As long as one is quoting a licence and the other a build, the offers aren't comparable.",[102,342,343],{},"Only once you have both of those does the \"dedicated or off-the-shelf\" conversation stop being a matter of opinion and become a calculation.",{"title":345,"searchDepth":346,"depth":346,"links":347},"",3,[348,355,356,357,358,359],{"id":116,"depth":349,"text":117,"children":350},2,[351,352,353,354],{"id":121,"depth":346,"text":122},{"id":128,"depth":346,"text":129},{"id":135,"depth":346,"text":136},{"id":142,"depth":346,"text":143},{"id":149,"depth":349,"text":150},{"id":257,"depth":349,"text":258},{"id":283,"depth":349,"text":284},{"id":300,"depth":349,"text":301},{"id":322,"depth":349,"text":323},"erp","2026-09-02","Decision criteria, a five-year cost comparison, and the signals that tell you an off-the-shelf ERP has stopped fitting your company.",false,null,"md","\u002Fimg\u002Fblog\u002Ferp-dedykowany-czy-gotowy.jpg",{},true,"\u002Fblog\u002Fdedykowany-erp-czy-gotowy-system","---\ntitle: Dedicated ERP or off-the-shelf — how to choose\ndescription: Decision criteria, a five-year cost comparison, and the signals that tell you an off-the-shelf ERP has stopped fitting your company.\ndate: 2026-09-02\ncategory: erp\ntags:\n  - erp\n  - rollouts\n  - decisions\nauthor: zespol\nimage: \u002Fimg\u002Fblog\u002Ferp-dedykowany-czy-gotowy.jpg\ntranslationKey: custom-erp-or-off-the-shelf\ndraft: false\n---\n\nThe answer is: **off-the-shelf, in most cases**. Worth stating up front, since a piece written by a company that builds dedicated systems might be expected to argue the opposite.\n\nAn off-the-shelf ERP wins when your process resembles other companies' in the industry and the budget is limited. Licence plus rollout is then cheaper than building, and the risk is lower. A dedicated system makes sense in a narrower set of situations — and this piece is about recognising when you're in that set.\n\n## Four deciding questions\n\n### 1. Is your process a source of advantage?\n\nIf a company makes money because it does something differently from its competitors, flattening that \"differently\" into a standard module costs more than the licence. If, on the other hand, the process is standard and the advantage lies elsewhere — relationships, price, location — an off-the-shelf system is the right choice.\n\n### 2. What does adapting the off-the-shelf system cost?\n\nThis question settles it most often. Ask the vendor for a quote on the modifications you need. If the sum of licence, rollout and modifications approaches the cost of building your own system, the choice stops being obvious — because for the same money you get either an 80%-fit system, or a 100%-fit one.\n\n### 3. How many users do you have?\n\nA per-user licence changes the maths over time. At twenty people the difference is negligible. At two hundred, over a five-year horizon, it can exceed the cost of building your own system.\n\n### 4. How much will the process change?\n\nA company in a stable industry, doing the same thing for a decade, fits well into an off-the-shelf system. A company changing its business model every two years will keep fighting limitations that can't be worked around except through the vendor.\n\n## The five-year comparison\n\nComparing only the rollout cost is misleading. Below is an example for a manufacturing company with 120 users. The figures are illustrative — meant to show the cost structure, not a quote.\n\n| Line item | Off-the-shelf ERP | Dedicated |\n|---|---|---|\n| Rollout \u002F build | PLN 250k | PLN 450k |\n| Licences over 5 years | PLN 430k | 0 |\n| Modifications and extensions | PLN 180k | PLN 150k |\n| Maintenance and support | PLN 120k | PLN 220k |\n| Infrastructure | included in licence | PLN 90k |\n| **Total** | **PLN 980k** | **PLN 910k** |\n\nThe takeaway isn't \"dedicated is cheaper.\" It's: **at this scale the costs converge, so the decision should be made on a different basis than price**. At thirty users the same table looks completely different, and the off-the-shelf system wins outright.\n\n## Signals that an off-the-shelf system has stopped fitting\n\n- The team keeps the \"real\" data in a spreadsheet next to the system.\n- A key report is produced by exporting to Excel and stitching it together by hand.\n- Every process change requires a ticket to the vendor and waits for quarters.\n- You're paying for modules nobody ever turned on.\n- A new hire is learning less the system than the list of workarounds the team invented.\n\nOne of these signals is normal friction. Three or more mean the cost of the mismatch has exceeded the cost of change.\n\n## A third path that's rarely discussed\n\nThe choice isn't binary. A common and underrated approach is to **keep an off-the-shelf ERP as the finance-and-accounting core and build your own modules where the company is genuinely atypical**.\n\nAccounting, HR and payroll are regulated areas where an in-house build is an expense with no return — the rules change every year and someone has to keep up. A production module, or handling a specific sales channel, is exactly where a custom fit pays off.\n\nThe condition is an ERP with a decent API. Without one, this path is closed — and that's an important criterion in choosing an off-the-shelf system that almost nobody asks about before signing.\n\n## What not to weigh in the decision\n\n**\"Everyone in the industry has X.\"** Your competitors made that decision at a different scale, in a different year, with a different process.\n\n**The sales demo.** Every system looks good in a demo the vendor prepared. Ask for access to a test environment and run your own real case through it — ideally the most unusual one you have.\n\n**The promise that \"it'll get configured.\"** Ask who will configure it, how long it'll take, and what a consultant's day costs. The answers are sometimes surprising.\n\n## How to make this decision properly\n\nDo two things before you choose.\n\nFirst: **write down your process as it actually is**, workarounds included. Those workarounds are information about what the previous tool was missing — and the best material for testing candidates.\n\nSecond: **ask both types of vendor to quote the same, written-down scope**. As long as one is quoting a licence and the other a build, the offers aren't comparable.\n\nOnly once you have both of those does the \"dedicated or off-the-shelf\" conversation stop being a matter of opinion and become a calculation.\n",{"title":96,"description":362},{"loc":369},"blog\u002Fdedykowany-erp-czy-gotowy-system",[360,375,376],"rollouts","decisions","custom-erp-or-off-the-shelf","SGZvpSbYxgHfAtuwEdBGzG6HA0KOjIg3dfXrHzqXe4U",{"id":380,"title":381,"author":97,"body":382,"category":651,"date":361,"description":652,"draft":363,"editorialNote":364,"extension":365,"image":653,"meta":654,"navigation":368,"path":655,"rawbody":656,"seo":657,"sitemap":658,"stem":659,"tags":660,"translationKey":664,"updated":364,"__hash__":665},"blogEn\u002Fblog\u002File-kosztuje-aplikacja-webowa.md","How much does a dedicated web application cost in 2026",{"type":99,"value":383,"toc":642},[384,395,398,402,463,466,470,473,505,509,515,521,527,533,539,545,551,555,558,583,587,597,600,604,610,616,622,628,632,639],[102,385,386,387,390,391,394],{},"A dedicated web application for a company costs, in Poland, anywhere from around ",[106,388,389],{},"PLN 60,000 net"," for a simple internal system to ",[106,392,393],{},"PLN 400,000 and up"," for a platform with multiple roles, integrations and a client portal. That spread isn't vendor indecision — it's that very different projects, differing by an order of magnitude in effort, hide under the same label.",[102,396,397],{},"Below I break those numbers down and show exactly what pushes a quote up.",[114,399,401],{"id":400},"price-ranges-by-project-type","Price ranges by project type",[155,403,404,417],{},[158,405,406],{},[161,407,408,411,414],{},[164,409,410],{},"Project type",[164,412,413],{},"Net range",[164,415,416],{},"Time to production",[174,418,419,430,441,452],{},[161,420,421,424,427],{},[179,422,423],{},"Internal tool, one process, one role",[179,425,426],{},"PLN 60–90k",[179,428,429],{},"3–4 months",[161,431,432,435,438],{},[179,433,434],{},"B2B system with a client portal and several roles",[179,436,437],{},"PLN 120–220k",[179,439,440],{},"5–8 months",[161,442,443,446,449],{},[179,444,445],{},"SaaS platform with subscriptions and multi-tenancy",[179,447,448],{},"PLN 250–400k",[179,450,451],{},"8–14 months",[161,453,454,457,460],{},[179,455,456],{},"Extending an existing system",[179,458,459],{},"PLN 15–60k per module",[179,461,462],{},"3–10 weeks",[102,464,465],{},"These are ranges for a Polish vendor with a senior team, at PLN 160–260 per hour. Offers significantly lower usually mean one of three things: a less experienced team, an underestimated scope, or a model where the gap gets made up in change orders.",[114,467,469],{"id":468},"where-the-money-actually-goes","Where the money actually goes",[102,471,472],{},"The cost breakdown for a typical mid-sized project looks like this:",[260,474,475,481,487,493,499],{},[263,476,477,480],{},[106,478,479],{},"Analysis and design — 15–20%."," Workshops, the data model, the prototype. The stage where it's easiest to save money and most expensive to have saved it.",[263,482,483,486],{},[106,484,485],{},"Backend and business logic — 35–40%."," Rules, permissions, integrations, performance.",[263,488,489,492],{},[106,490,491],{},"Interface — 25–30%."," The more roles and views, the more it costs. The admin panel is sometimes bigger than the client-facing part.",[263,494,495,498],{},[106,496,497],{},"Testing and rollout — 10–15%."," Environments, release automation, data migration.",[263,500,501,504],{},[106,502,503],{},"Buffer for the unknown — 10%."," Every project has one. Vendors who don't budget for it make it up in change orders instead.",[114,506,508],{"id":507},"seven-things-that-double-a-quote","Seven things that double a quote",[102,510,511,514],{},[106,512,513],{},"1. Integrations with systems that have no API."," Connecting to a well-documented API takes a few days. Connecting to a system that only offers a file export or direct database access takes weeks — and it's a debt that comes back with every update on that system's side.",[102,516,517,520],{},[106,518,519],{},"2. Migrating data from an inconsistent source."," If a spreadsheet has the same company appearing as \"Smith Ltd\", \"Smith Ltd.\" and \"SMITH LTD\", someone has to resolve that. There's usually more of it than it looks.",[102,522,523,526],{},[106,524,525],{},"3. Offline-first work."," An app that has to work with no connection and sync afterwards needs a local database, an operation queue and conflict-resolution rules. That can turn out to be a third of the whole project.",[102,528,529,532],{},[106,530,531],{},"4. Multilingual support bolted on later."," Adding a second language to a system that wasn't built for it costs several times more than planning for it from the start.",[102,534,535,538],{},[106,536,537],{},"5. A complex permissions model."," The gap between \"admin and user\" and a permissions matrix per module, per field and per branch is weeks of work and a standing cost on every feature after that.",[102,540,541,544],{},[106,542,543],{},"6. Compliance requirements."," Baseline GDPR is cheap. A requirement to audit every data change, retention rules and on-demand anonymisation is a module of its own.",[102,546,547,550],{},[106,548,549],{},"7. An open-ended scope."," The single biggest factor. A project where the scope gets settled along the way ends up costing, in practice, 1.5–2x a project with the same scope described up front.",[114,552,554],{"id":553},"what-a-quote-usually-doesnt-include","What a quote usually doesn't include",[102,556,557],{},"Vendors quote the cost of the build. You still need to budget for:",[260,559,560,566,571,577],{},[263,561,562,565],{},[106,563,564],{},"Maintenance"," — realistically 15–20% of the project's value per year. Monitoring, security updates, small fixes.",[263,567,568,570],{},[106,569,224],{}," — from a few hundred zloty a month for a simple system to several thousand at higher traffic.",[263,572,573,576],{},[106,574,575],{},"Content and photography"," — negligible for internal systems, potentially a significant line item for products with a public-facing part.",[263,578,579,582],{},[106,580,581],{},"Your team's time"," — workshops, acceptance testing, decisions. A few to a dozen hours a week across the whole project. It's a real cost, just not one that shows up on an invoice.",[114,584,586],{"id":585},"how-to-check-whether-a-quote-is-fair","How to check whether a quote is fair",[102,588,589,590,593,594,291],{},"Ask for a breakdown by stage with amounts, and a list of assumptions. A fair quote states not just what's in scope, but also ",[106,591,592],{},"what isn't",", and ",[106,595,596],{},"what happens to the price if an assumption doesn't hold",[102,598,599],{},"If you get a single number with no breakdown, you have no way to compare it to another single number. Three offers differing by an order of magnitude almost always mean each of them priced something different.",[114,601,603],{"id":602},"how-to-lower-the-cost-without-lowering-quality","How to lower the cost without lowering quality",[102,605,606,609],{},[106,607,608],{},"Narrow the first rollout to a single process."," A system that handles one process end to end delivers value from day one. A system half-handling five processes delivers none.",[102,611,612,615],{},[106,613,614],{},"Skip configurability you don't need yet."," \"We want to be able to change this later without a developer\" is usually the most expensive sentence in the whole spec.",[102,617,618,621],{},[106,619,620],{},"Use off-the-shelf components where you're not building an advantage."," Authentication, payments, email delivery, search — good components already exist, and writing them from scratch is a cost with no return.",[102,623,624,627],{},[106,625,626],{},"Pay for a workshop before building."," A dozen or so thousand spent closing the scope pays for itself many times over, because it removes the main cause of budget overruns.",[114,629,631],{"id":630},"summary","Summary",[102,633,634,635,638],{},"If you need one number to open a conversation with the board: ",[106,636,637],{},"an internal system handling one process is around PLN 80,000 net and four months",". Everything above that depends on the number of roles, integrations, and how well the scope is described — not on how many screens the app has.",[102,640,641],{},"If you want a number you can defend to the board, you need a workshop on your own data and your own process. That's the only way a quote stops being a guess.",{"title":345,"searchDepth":346,"depth":346,"links":643},[644,645,646,647,648,649,650],{"id":400,"depth":349,"text":401},{"id":468,"depth":349,"text":469},{"id":507,"depth":349,"text":508},{"id":553,"depth":349,"text":554},{"id":585,"depth":349,"text":586},{"id":602,"depth":349,"text":603},{"id":630,"depth":349,"text":631},"aplikacje-webowe","Real price ranges for web applications, a cost breakdown by stage, and the factors that can double a quote. No \"it depends\" without an explanation.","\u002Fimg\u002Fblog\u002File-kosztuje-aplikacja-webowa.jpg",{},"\u002Fblog\u002File-kosztuje-aplikacja-webowa","---\ntitle: How much does a dedicated web application cost in 2026\ndescription: Real price ranges for web applications, a cost breakdown by stage, and the factors that can double a quote. No \"it depends\" without an explanation.\ndate: 2026-09-02\ncategory: aplikacje-webowe\ntags:\n  - pricing\n  - budget\n  - mvp\nauthor: zespol\nimage: \u002Fimg\u002Fblog\u002File-kosztuje-aplikacja-webowa.jpg\ntranslationKey: cost-of-a-web-application\ndraft: false\n---\n\nA dedicated web application for a company costs, in Poland, anywhere from around **PLN 60,000 net** for a simple internal system to **PLN 400,000 and up** for a platform with multiple roles, integrations and a client portal. That spread isn't vendor indecision — it's that very different projects, differing by an order of magnitude in effort, hide under the same label.\n\nBelow I break those numbers down and show exactly what pushes a quote up.\n\n## Price ranges by project type\n\n| Project type | Net range | Time to production |\n|---|---|---|\n| Internal tool, one process, one role | PLN 60–90k | 3–4 months |\n| B2B system with a client portal and several roles | PLN 120–220k | 5–8 months |\n| SaaS platform with subscriptions and multi-tenancy | PLN 250–400k | 8–14 months |\n| Extending an existing system | PLN 15–60k per module | 3–10 weeks |\n\nThese are ranges for a Polish vendor with a senior team, at PLN 160–260 per hour. Offers significantly lower usually mean one of three things: a less experienced team, an underestimated scope, or a model where the gap gets made up in change orders.\n\n## Where the money actually goes\n\nThe cost breakdown for a typical mid-sized project looks like this:\n\n- **Analysis and design — 15–20%.** Workshops, the data model, the prototype. The stage where it's easiest to save money and most expensive to have saved it.\n- **Backend and business logic — 35–40%.** Rules, permissions, integrations, performance.\n- **Interface — 25–30%.** The more roles and views, the more it costs. The admin panel is sometimes bigger than the client-facing part.\n- **Testing and rollout — 10–15%.** Environments, release automation, data migration.\n- **Buffer for the unknown — 10%.** Every project has one. Vendors who don't budget for it make it up in change orders instead.\n\n## Seven things that double a quote\n\n**1. Integrations with systems that have no API.** Connecting to a well-documented API takes a few days. Connecting to a system that only offers a file export or direct database access takes weeks — and it's a debt that comes back with every update on that system's side.\n\n**2. Migrating data from an inconsistent source.** If a spreadsheet has the same company appearing as \"Smith Ltd\", \"Smith Ltd.\" and \"SMITH LTD\", someone has to resolve that. There's usually more of it than it looks.\n\n**3. Offline-first work.** An app that has to work with no connection and sync afterwards needs a local database, an operation queue and conflict-resolution rules. That can turn out to be a third of the whole project.\n\n**4. Multilingual support bolted on later.** Adding a second language to a system that wasn't built for it costs several times more than planning for it from the start.\n\n**5. A complex permissions model.** The gap between \"admin and user\" and a permissions matrix per module, per field and per branch is weeks of work and a standing cost on every feature after that.\n\n**6. Compliance requirements.** Baseline GDPR is cheap. A requirement to audit every data change, retention rules and on-demand anonymisation is a module of its own.\n\n**7. An open-ended scope.** The single biggest factor. A project where the scope gets settled along the way ends up costing, in practice, 1.5–2x a project with the same scope described up front.\n\n## What a quote usually doesn't include\n\nVendors quote the cost of the build. You still need to budget for:\n\n- **Maintenance** — realistically 15–20% of the project's value per year. Monitoring, security updates, small fixes.\n- **Infrastructure** — from a few hundred zloty a month for a simple system to several thousand at higher traffic.\n- **Content and photography** — negligible for internal systems, potentially a significant line item for products with a public-facing part.\n- **Your team's time** — workshops, acceptance testing, decisions. A few to a dozen hours a week across the whole project. It's a real cost, just not one that shows up on an invoice.\n\n## How to check whether a quote is fair\n\nAsk for a breakdown by stage with amounts, and a list of assumptions. A fair quote states not just what's in scope, but also **what isn't**, and **what happens to the price if an assumption doesn't hold**.\n\nIf you get a single number with no breakdown, you have no way to compare it to another single number. Three offers differing by an order of magnitude almost always mean each of them priced something different.\n\n## How to lower the cost without lowering quality\n\n**Narrow the first rollout to a single process.** A system that handles one process end to end delivers value from day one. A system half-handling five processes delivers none.\n\n**Skip configurability you don't need yet.** \"We want to be able to change this later without a developer\" is usually the most expensive sentence in the whole spec.\n\n**Use off-the-shelf components where you're not building an advantage.** Authentication, payments, email delivery, search — good components already exist, and writing them from scratch is a cost with no return.\n\n**Pay for a workshop before building.** A dozen or so thousand spent closing the scope pays for itself many times over, because it removes the main cause of budget overruns.\n\n## Summary\n\nIf you need one number to open a conversation with the board: **an internal system handling one process is around PLN 80,000 net and four months**. Everything above that depends on the number of roles, integrations, and how well the scope is described — not on how many screens the app has.\n\nIf you want a number you can defend to the board, you need a workshop on your own data and your own process. That's the only way a quote stops being a guess.\n",{"title":381,"description":652},{"loc":655},"blog\u002File-kosztuje-aplikacja-webowa",[661,662,663],"pricing","budget","mvp","cost-of-a-web-application","jhCI0ZJN6yMJqQD2Jbr2XiEVwicPPV8synd3rKHAvg4",{"id":667,"title":668,"author":97,"body":669,"category":880,"date":361,"description":881,"draft":363,"editorialNote":364,"extension":365,"image":882,"meta":883,"navigation":368,"path":884,"rawbody":885,"seo":886,"sitemap":887,"stem":888,"tags":889,"translationKey":891,"updated":364,"__hash__":892},"blogEn\u002Fblog\u002Fwdrozenie-ai-w-firmie-od-czego-zaczac.md","AI adoption in your company — where to start",{"type":99,"value":670,"toc":870},[671,685,689,692,699,703,706,712,718,724,727,731,738,741,745,748,754,760,766,772,776,779,799,802,806,832,836,842,848,854,860,864,867],[102,672,673,674,677,678,593,681,684],{},"Start with a process that meets three conditions at once: it's ",[106,675,676],{},"repetitive",", has a ",[106,679,680],{},"countable time cost",[106,682,683],{},"tolerates a mistake",". Classifying incoming documents usually meets all three. \"An assistant that knows the whole company\" meets none of them — which is why so many rollouts that start there never finish.",[114,686,688],{"id":687},"why-pilots-dont-reach-production","Why pilots don't reach production",[102,690,691],{},"A demo shows a model handling a dozen or so examples chosen by the person who prepared them. Production is tens of thousands of examples nobody chose: scans crooked on the glass, documents in a format outside the test set, invoices with a handwritten note in the margin.",[102,693,694,695,698],{},"The difference between the two isn't model quality. It's ",[106,696,697],{},"handling the cases where the model isn't sure"," — and that's exactly what a rollout is actually made of. A pilot that skips it isn't an earlier version of the rollout. It's a different thing entirely.",[114,700,702],{"id":701},"how-to-choose-your-first-use-case","How to choose your first use case",[102,704,705],{},"Make a list of candidates and score each on three dimensions.",[102,707,708,711],{},[106,709,710],{},"Repeatability."," How many times a month does someone do this? Below a hundred, the savings rarely justify a rollout.",[102,713,714,717],{},[106,715,716],{},"Time cost."," How many minutes does one instance take? Multiply by repeatability. That number sets the upper bound on a sensible budget.",[102,719,720,723],{},[106,721,722],{},"Tolerance for error."," What happens if the model gets it wrong? If the answer is \"someone catches it at review\" — good. If it's \"money leaves an account\" — you need a human approval step, which changes the maths.",[102,725,726],{},"The best candidates in a typical company: classifying and routing incoming documents, extracting data from invoices and orders, drafting first responses to repetitive support questions, searching internal documentation.",[114,728,730],{"id":729},"set-a-success-metric-before-you-start","Set a success metric before you start",[102,732,733,734],{},"Before a prototype exists, write down one sentence: ",[735,736,737],"em",{},"\"This rollout makes sense if the model handles at least X% of cases on its own, at accuracy of at least Y%.\"",[102,739,740],{},"Without that number, there's no way to later decide whether the project succeeded. With it, the prototype ends in a clear decision: proceed, or stop. Stopping is also a good outcome — just a cheaper one than finding out a year in.",[114,742,744],{"id":743},"an-architecture-that-survives-production","An architecture that survives production",[102,746,747],{},"A rollout that lasts has four elements a demo doesn't.",[102,749,750,753],{},[106,751,752],{},"A confidence threshold."," The model returns a result along with its own confidence score. Above the threshold, automation runs; below it, the case goes to a human. Setting that threshold is a business decision, not a technical one.",[102,755,756,759],{},[106,757,758],{},"A queue for uncertain cases."," The place where a human resolves cases below the threshold — and its size is the real measure of how well the system is working.",[102,761,762,765],{},[106,763,764],{},"A decision log."," A record of what the model received, what it returned, and with what confidence. Without it you can't diagnose degradation or answer an auditor's question.",[102,767,768,771],{},[106,769,770],{},"Continuous measurement."," Data changes. A model that was 94% accurate in March can be at 80% by November — not because it broke, but because the input stream changed.",[114,773,775],{"id":774},"where-your-data-goes","Where your data goes",[102,777,778],{},"The question that comes up in every first meeting, rightly so. You have three options:",[260,780,781,787,793],{},[263,782,783,786],{},[106,784,785],{},"A hosted model, no-training agreement."," Data leaves the company, but the provider commits not to use it for training. Cheapest and fastest, sufficient for most use cases.",[263,788,789,792],{},[106,790,791],{},"A hosted model in your own private cloud."," Data never leaves your infrastructure. More expensive, justified for sensitive data.",[263,794,795,798],{},[106,796,797],{},"A model on your own hardware."," Full control, the highest entry and running cost. Makes sense under regulatory requirements that rule out the other options.",[102,800,801],{},"The choice is a decision about risk, not technology. You make it once, and it affects cost for the entire lifetime of the solution.",[114,803,805],{"id":804},"what-it-costs","What it costs",[260,807,808,814,820,826],{},[263,809,810,813],{},[106,811,812],{},"Prototype on your data:"," PLN 15–30k net, 2–3 weeks. Ends with a number you can base a decision on.",[263,815,816,819],{},[106,817,818],{},"Production rollout:"," PLN 60–200k net, depending on the number of integrations and approval requirements.",[263,821,822,825],{},[106,823,824],{},"Running the models:"," from a few hundred to a few thousand zloty a month, roughly proportional to volume.",[263,827,828,831],{},[106,829,830],{},"Maintenance and tuning:"," usually 15% of the rollout's value per year.",[114,833,835],{"id":834},"four-mistakes-that-come-up-most-often","Four mistakes that come up most often",[102,837,838,841],{},[106,839,840],{},"Starting with the most impressive use case."," A conversational assistant impresses the board and is the hardest possible starting point — unbounded scope, no good success metric, high expectations.",[102,843,844,847],{},[106,845,846],{},"No human in the loop."," A system with no path for uncertain cases either performs badly or needs such a high confidence threshold that it only automates a small share of cases.",[102,849,850,853],{},[106,851,852],{},"Measuring accuracy alone."," 95% accuracy sounds great until you work out that at a thousand documents a day, that's fifty errors someone has to catch.",[102,855,856,859],{},[106,857,858],{},"Treating the rollout as a closed project."," A model needs watching, like any other piece of production infrastructure. A rollout with no measurement plan ages quietly.",[114,861,863],{"id":862},"the-first-step-you-can-take-this-week","The first step you can take this week",[102,865,866],{},"Pick one process. Count how many times a month it happens and how many minutes it takes. Write down what happens when someone gets it wrong.",[102,868,869],{},"Those three numbers are enough to judge whether a prototype is worth building — and they're worth more than a month of conversations about AI's possibilities.",{"title":345,"searchDepth":346,"depth":346,"links":871},[872,873,874,875,876,877,878,879],{"id":687,"depth":349,"text":688},{"id":701,"depth":349,"text":702},{"id":729,"depth":349,"text":730},{"id":743,"depth":349,"text":744},{"id":774,"depth":349,"text":775},{"id":804,"depth":349,"text":805},{"id":834,"depth":349,"text":835},{"id":862,"depth":349,"text":863},"ai","How to choose your first use case, what it costs, and why most corporate AI projects end at the demo instead of reaching production.","\u002Fimg\u002Fblog\u002Fwdrozenie-ai-w-firmie.jpg",{},"\u002Fblog\u002Fwdrozenie-ai-w-firmie-od-czego-zaczac","---\ntitle: AI adoption in your company — where to start\ndescription: How to choose your first use case, what it costs, and why most corporate AI projects end at the demo instead of reaching production.\ndate: 2026-09-02\ncategory: ai\ntags:\n  - ai\n  - automation\n  - rollouts\nauthor: zespol\nimage: \u002Fimg\u002Fblog\u002Fwdrozenie-ai-w-firmie.jpg\ntranslationKey: ai-adoption-where-to-start\ndraft: false\n---\n\nStart with a process that meets three conditions at once: it's **repetitive**, has a **countable time cost**, and **tolerates a mistake**. Classifying incoming documents usually meets all three. \"An assistant that knows the whole company\" meets none of them — which is why so many rollouts that start there never finish.\n\n## Why pilots don't reach production\n\nA demo shows a model handling a dozen or so examples chosen by the person who prepared them. Production is tens of thousands of examples nobody chose: scans crooked on the glass, documents in a format outside the test set, invoices with a handwritten note in the margin.\n\nThe difference between the two isn't model quality. It's **handling the cases where the model isn't sure** — and that's exactly what a rollout is actually made of. A pilot that skips it isn't an earlier version of the rollout. It's a different thing entirely.\n\n## How to choose your first use case\n\nMake a list of candidates and score each on three dimensions.\n\n**Repeatability.** How many times a month does someone do this? Below a hundred, the savings rarely justify a rollout.\n\n**Time cost.** How many minutes does one instance take? Multiply by repeatability. That number sets the upper bound on a sensible budget.\n\n**Tolerance for error.** What happens if the model gets it wrong? If the answer is \"someone catches it at review\" — good. If it's \"money leaves an account\" — you need a human approval step, which changes the maths.\n\nThe best candidates in a typical company: classifying and routing incoming documents, extracting data from invoices and orders, drafting first responses to repetitive support questions, searching internal documentation.\n\n## Set a success metric before you start\n\nBefore a prototype exists, write down one sentence: *\"This rollout makes sense if the model handles at least X% of cases on its own, at accuracy of at least Y%.\"*\n\nWithout that number, there's no way to later decide whether the project succeeded. With it, the prototype ends in a clear decision: proceed, or stop. Stopping is also a good outcome — just a cheaper one than finding out a year in.\n\n## An architecture that survives production\n\nA rollout that lasts has four elements a demo doesn't.\n\n**A confidence threshold.** The model returns a result along with its own confidence score. Above the threshold, automation runs; below it, the case goes to a human. Setting that threshold is a business decision, not a technical one.\n\n**A queue for uncertain cases.** The place where a human resolves cases below the threshold — and its size is the real measure of how well the system is working.\n\n**A decision log.** A record of what the model received, what it returned, and with what confidence. Without it you can't diagnose degradation or answer an auditor's question.\n\n**Continuous measurement.** Data changes. A model that was 94% accurate in March can be at 80% by November — not because it broke, but because the input stream changed.\n\n## Where your data goes\n\nThe question that comes up in every first meeting, rightly so. You have three options:\n\n- **A hosted model, no-training agreement.** Data leaves the company, but the provider commits not to use it for training. Cheapest and fastest, sufficient for most use cases.\n- **A hosted model in your own private cloud.** Data never leaves your infrastructure. More expensive, justified for sensitive data.\n- **A model on your own hardware.** Full control, the highest entry and running cost. Makes sense under regulatory requirements that rule out the other options.\n\nThe choice is a decision about risk, not technology. You make it once, and it affects cost for the entire lifetime of the solution.\n\n## What it costs\n\n- **Prototype on your data:** PLN 15–30k net, 2–3 weeks. Ends with a number you can base a decision on.\n- **Production rollout:** PLN 60–200k net, depending on the number of integrations and approval requirements.\n- **Running the models:** from a few hundred to a few thousand zloty a month, roughly proportional to volume.\n- **Maintenance and tuning:** usually 15% of the rollout's value per year.\n\n## Four mistakes that come up most often\n\n**Starting with the most impressive use case.** A conversational assistant impresses the board and is the hardest possible starting point — unbounded scope, no good success metric, high expectations.\n\n**No human in the loop.** A system with no path for uncertain cases either performs badly or needs such a high confidence threshold that it only automates a small share of cases.\n\n**Measuring accuracy alone.** 95% accuracy sounds great until you work out that at a thousand documents a day, that's fifty errors someone has to catch.\n\n**Treating the rollout as a closed project.** A model needs watching, like any other piece of production infrastructure. A rollout with no measurement plan ages quietly.\n\n## The first step you can take this week\n\nPick one process. Count how many times a month it happens and how many minutes it takes. Write down what happens when someone gets it wrong.\n\nThose three numbers are enough to judge whether a prototype is worth building — and they're worth more than a month of conversations about AI's possibilities.\n",{"title":668,"description":881},{"loc":884},"blog\u002Fwdrozenie-ai-w-firmie-od-czego-zaczac",[880,890,375],"automation","ai-adoption-where-to-start","3h3Uyc_r6v-R1oggm3sXVR3s8arJjyy0xQy78uryQSg",1790522540631]