Why Agencies Charge Before Building Anything
You describe your idea to a software agency. The conversation goes well. Then the proposal arrives, and the first item on the list — before any code, before any design, before anything tangible — is a discovery phase. Price: $8,000 to $15,000. Timeline: three to four weeks.
It is a common moment of friction. You came to get something built. They are asking you to pay them to figure out what to build. It feels like being charged twice for the same thing.
It isn't. The discovery phase and the build phase are solving different problems. The build phase constructs the software. The discovery phase defines what the software actually is, what it needs to connect to, what edge cases exist, and what 'done' means. Without that definition, you are not getting a fixed-price project — you are getting an estimate based on assumptions that will be revised as soon as real work begins.
What Actually Happens During Discovery
A properly run discovery phase is not a sales exercise or a stalling tactic. It is a diagnostic. The agency is doing the work required to give you an accurate quote and a credible timeline — work that cannot be done from a one-hour initial call.
In practical terms, this typically includes a thorough review of your existing systems, data, and workflows; structured interviews with the people in your business who will actually use the software; a technical audit of any third-party integrations the new system will need to connect to; user story mapping that translates business requirements into defined acceptance criteria; and an architecture plan with explicit decisions about technology stack, infrastructure, and data flows.
The output is not a polished presentation. It is a working document: a detailed backlog with realistic estimates, a phased roadmap, a risk register for known unknowns, and a definition of what production-ready looks like for your specific project. A serious agency will put a fixed number next to phase two of your project only after this work is done — because before it is done, any number is a guess.
Key Takeaways
- Discovery reviews existing systems, data, integrations, and user workflows — not just the idea
- Output includes a detailed backlog with estimates, a phased roadmap, and explicit acceptance criteria
- A fixed-price build quote produced without discovery is a guess, not a commitment
- Typical discovery takes 2–4 weeks and costs 5–15% of the total estimated project budget
The Real Cost of Skipping It
The most commonly cited data point in this conversation is stark: companies that skip a structured discovery phase spend 40–60% more on development than those that invest in it upfront. That figure is not derived from unusual projects or bad actors. It reflects the normal cost of building software based on assumptions rather than verified requirements.
Here is how the pattern typically unfolds. You hire an agency that skips discovery and quotes you a fast, attractive number. Work begins. Three weeks in, it becomes clear that the integration with your CRM is more complex than expected — that is a change order. Four weeks later, the edge case you never mentioned (because no one asked) requires a significant rework of the data model — another change order. The budget climbs. The timeline slips. The original quote, which felt like a deal, ends up looking like the setup for a much larger invoice.
The agencies that do this are not universally dishonest. Many are simply skipping the work they should have done before quoting — whether because it speeds up the sales cycle or because they genuinely do not know how to scope work properly. The outcome for you is the same either way.
A discovery investment of $10,000 that prevents $40,000 in downstream scope changes is not an added cost. It is a return. The businesses that recognize this are the ones that tend to end projects on budget.
Key Takeaways
- Skipping discovery correlates with 40–60% higher final development costs
- Change orders are the mechanism: undefined scope becomes priced work after you have committed
- Discovery investment of 5–10% of project budget typically prevents overruns of 40–60%
- Both inexperienced and incentive-misaligned vendors skip discovery — the outcome is the same for you
What a Legitimate Discovery Phase Produces
Not every agency that charges for discovery is running a legitimate process. The way to tell the difference is to ask what you will receive at the end — specifically, and in writing.
A real discovery deliverable includes: a documented requirements specification tied to your actual workflows, not a generic template; an architecture diagram showing how the new system connects to your existing tools; a backlog of user stories with acceptance criteria and effort estimates; a phased roadmap that distinguishes must-have from nice-to-have; and a risk register that names the specific unknowns the team has identified and how they plan to address them.
If an agency cannot tell you what you will receive at the end of discovery before you pay for it, they are not running a discovery process — they are buying time. The deliverable list should be in the contract. The definition of what 'done' means for the discovery phase itself should be as clear as the definition of what 'done' means for the build.
Key Takeaways
- Ask for the discovery deliverables list before you sign — it should be in the contract
- Architecture diagram, requirements spec, backlog with estimates, phased roadmap, and risk register are the baseline
- A vague answer to 'what will I receive at the end?' is a red flag, not a reassurance
When to Push Back — and When Not To
Discovery is not always warranted at the same depth. A small, well-defined project with minimal integrations and a clear scope may need a lighter-weight scoping session rather than a multi-week engagement. If an agency is proposing a $15,000 discovery phase for a project you expect to cost $20,000, that proportion deserves scrutiny.
The cases where discovery is clearly worth the investment are: any project involving integrations with multiple existing systems; any project where your internal data is scattered or poorly structured; any project where the business requirements are not yet fully defined; and any project over approximately $30,000 in estimated build cost. In these cases, the discovery phase is not optional — it is the mechanism that makes everything downstream reliable.
The right way to evaluate any discovery proposal is not 'do I want to pay for this?' but 'what happens if we skip it?' If the honest answer is 'we proceed on assumptions and handle the consequences as change orders,' that is a useful framing. Discovery transfers the uncertainty from the invoicing stage to the scoping stage — where it is much cheaper to resolve.
A development partner who recommends discovery before recommending a specific technology or a fixed price is giving you a signal worth paying attention to. In a market where many agencies are optimizing for fast contract signatures rather than successful projects, a willingness to slow down and define the work before pricing it is one of the more reliable indicators that you are dealing with a team that has actually delivered software — not just sold it.
The Bottom Line
The discovery phase question is, at its core, a question about where you want risk to live in your project. Keep it in the build phase — where it materializes as change orders, budget overruns, and rework — or surface it in the scoping phase, where it is far cheaper to address. At StepTo, discovery is the first thing we do for any meaningful build engagement. We assess your data, map your integrations, and produce a detailed plan with realistic estimates before we quote you a build price. If you have a proposal in hand and you are not sure whether the scoping process behind it is legitimate — or if you want to start a project with a partner that takes discovery seriously — that is exactly the conversation to start with us.
Building a team in Eastern Europe?
StepTo helps European and US companies build senior-led nearshore engineering teams in Serbia. Let's talk about what your next engagement could look like.
Start a conversation