Choosing a Web Agency in Novi Sad - A Scoring Framework

When a business owner in Novi Sad starts looking for a web agency, the first search usually returns fifteen or so results. Portfolios look similar, prices are not comparable, everyone promises fast, high-quality and cheap. After three conversations, the decision gets made on gut feel - or based on who replied first. The problem is that gut feel does not build a good website. A clear selection method does. In this post we lay out the scorecard with 10 criteria that we use and recommend to clients - the same method you should run three agencies through before signing a contract.

The short answer - why a scorecard beats gut feel

A scorecard is a simple table in which you grade each agency from 1 to 5 across ten predefined criteria, each with its own weight. At the end you have three numbers you can compare. It sounds overly formal for picking a website vendor, but the difference in the final decision is huge - instead of picking the agency you liked best in conversation, you pick the agency that objectively fits the project best.

The method is especially useful in Novi Sad and Belgrade, where the market has a wide spread of agency profiles - freelancers, small studios of three to five people, mid-sized agencies of ten to twenty, and large studios. Without a framework, it is hard to compare a solo developer's offer with a ten-person team's offer.

10 scorecard criteria with weights

Each agency you are considering gets a 1-to-5 score across the following criteria. The weight shows the multiplier - the higher the weight, the more important the criterion. The maximum total is 150 points.

• Portfolio with live links - weight 5 (open 5-10 sites, check speed and mobile view)
• Concrete results from previous projects - weight 4 (rankings, conversions, traffic, not just visuals)
• Clarity of communication in the first call - weight 4 (direct answers, no empty promises)
• Technology they use - weight 3 (modern, maintainable for the next 3-5 years)
• Price transparency - weight 3 (fixed figure, detailed scope, no hidden items)
• Ownership of code and access - weight 3 (everything is yours after delivery)
• References you can actually call - weight 2 (two or three past clients ready to talk)
• Team visible on the site - weight 2 (names, photos, LinkedIn profiles, not stock images)
• Post-launch support model - weight 2 (a clear maintenance offer, not 'we will see')
• Local context - weight 2 (understanding of the Serbian market and google.rs rankings)

When you multiply the score by the weight and add it all up, you get a concrete number between 30 and 150. An agency that scores above 110 is a serious candidate. One below 85 usually is not. The band in between is where you look at additional factors - gut feel, chemistry, availability in your timeline.

Phase 1 - narrowing 15 agencies down to 5

The first phase runs without any calls, purely from the agency's website and portfolio. The goal is to narrow fifteen candidates found in search down to five worth contacting. The pre-selection takes no longer than an hour total, and it eliminates eighty percent of candidates who would otherwise waste your time later.

• Open their site - does it load fast and look professional?
• Open the portfolio - does it exist at all, and does it include live links from the last two years?
• Open three sites from the portfolio - do they work, do they look good on mobile?
• Check the 'About' page - is there a concrete team or just stock photos?
• Check the 'Contact' page - is there a physical address, phone number, company tax ID?
• Check the pricing page or at least rough packages - is there any transparency?

Agencies that fail two or three of these checks get cut immediately. For the rest, before you call them, prepare a short brief - what you need, rough budget and timeline. We covered how to write that document in detail in our post on preparing before you call an agency.

Phase 2 - the first call and narrowing down to 3

The five remaining agencies each get a 30 to 45 minute call. The goal is not to get a quote but to fill in the scorecard against criteria 3 through 10 (you already scored 1 and 2 from the portfolio). Do not run five calls in the same week - memory fades, impressions blur. Two calls per day, maximum.

The questions you ask have to be the same across all agencies, so that the answers are comparable. The minimum set of questions that reveals most of what you need to know fits into these ten.

• Who owns the code and design after delivery - all mine, or do you retain any rights?
• What technology do you use and why that one for my type of project?
• How many similar projects have you done and can I see at least two?
• What is included in the price, and what gets billed separately?
• What is the rough timeline and what happens if you overrun it?
• What does support after launch look like - monthly retainer, hourly, packages?
• Do you include basic SEO optimization and which elements specifically?
• How many design iterations and content revisions are included?
• Do we sign a contract before I pay a deposit, and what are the termination terms?
• What is the payment schedule - how much upfront, how much in phases, how much at the end?

After each call, within twenty minutes while the conversation is still fresh, fill in the scorecard. Do not wait until evening - scores that have faded from memory are no longer accurate. The three agencies with the highest totals move to the next phase.

Phase 3 - detailed proposal and the final decision

The three finalists are asked for a detailed written proposal with the page scope, functionality, timeline, payment schedule and contract terms. A serious agency delivers that proposal within five to seven business days. If one is late without explanation, that itself is information - it means they will be late during the project too.

Once all three proposals are in, compare them across the following parameters - not just price:

• Clarity of scope - how detailed is the description of the functionality?
• Fixed price vs range - where does an 'estimate' live, and what moves the number up?
• Deposit vs delivery split - 40-50% upfront is standard, more than 70% is a risk
• Quality guarantee - what happens if the site crashes, has bugs, or fails to hit agreed performance?
• Ownership rights - is it clearly written who owns what after delivery?
• Timeline with milestones - not just 'done in 6 weeks' but clear intermediate steps

The lowest offer is almost never the best deal. If one agency offers half the price of the others, there is usually a reason - a template instead of custom work, vague scope, add-ons that surface during the project. Concrete example: a 500€ offer that does not include copy, photos, SEO and hosting ends up at 1500€ by the end of the project.

How to adapt the scorecard to your business type

The ten criteria work for most projects, but the weights change depending on what you are building. Here are three common scenarios we see in Novi Sad and Belgrade.

Local service business (dentist, lawyer, car wash, restaurant) - raise the weights of 'local context' and 'concrete results from previous projects' from 2 and 4 to 4 and 5. For this type of business, it is critical that the site ranks on google.rs for searches like 'dentist Novi Sad' or 'lawyer Belgrade'. An agency that does not understand local ranking can build a beautiful site that nobody finds.

Online store - raise 'technology' and 'support model' from 3 and 2 to 5 and 4. A store is a living system that requires payment, shipping and accounting integrations, plus ongoing support when something breaks. We compared platforms in detail in our Shopify, WooCommerce and custom comparison.

Custom web application (internal tool, SaaS platform, booking system) - raise 'technology' and 'portfolio with live links' to maximum weight. An application demands serious architecture and a team that has shipped similar projects. A freelancer who builds good WordPress sites cannot safely build a SaaS platform. We covered the difference between a site and an app on our application development service page.

The local angle - working with a Novi Sad agency

Over the past five years Novi Sad has grown into one of the stronger IT hubs in the region, and the number of agencies focused on web development has doubled. For a business in Vojvodina or Belgrade, working with a Novi Sad agency brings a few concrete advantages - the ability to meet in person when a complex topic needs quick resolution, communication without a language barrier, and familiarity with Serbian laws and payment systems. Prices in Novi Sad are usually more competitive than in Belgrade thanks to lower overhead, while the quality of work is identical.

This does not mean every Novi Sad agency delivers - it just means you have a solid foundation for a long-term partnership. The scorecard still has to be passed by every candidate, regardless of location.

Common scoring mistakes

The method only works if applied consistently. Here is where clients tend to go wrong and how it leads to the wrong decision.

• Giving a high score to a 'charming salesperson' - the salesperson does not build the site, the team behind them does; make sure you will work directly with that team
• Overweighting price as a criterion - price is kept out of the scorecard on purpose, it is evaluated separately at the end; do not sneak it in as a hidden factor
• Filling in the scorecard a week after the call - impressions fade, fill it in immediately
• Giving a score of 3 when you do not know - a 3 devalues the system; better to call the agency back and ask specifically
• Ignoring red flags because the score is high - if there is a serious red flag (no contract, no code ownership), it is an automatic disqualification regardless of score
• Adding an eleventh criterion mid-process - the system works because it is defined upfront; changing it mid-way breaks comparability

Conclusion

Choosing a web agency in Novi Sad does not have to be a game of chance. A scorecard with 10 criteria, three selection phases and clearly defined questions give you a concrete method for comparing five to fifteen candidates and picking the one that objectively fits the project best. The method is not slow - the full pre-selection fits into an hour, and the difference between a site that brings clients and a site that does nothing often fits into that one hour invested at the start.

If you want to see how we apply these criteria in our own work, take a look at our pricing or reach out through our contact page - in the first conversation you will get direct answers to all ten scorecard questions, no deflection and no hidden items.