WhatsApp
Chat on WhatsApp

Systems Analysis: How to Spec a System Before You Build

Almost every business today runs several systems side by side: a website, a CRM, an email marketing tool, a payment provider, accounting software and often a few automations that tie them together. As long as everything works, nobody thinks about those connections. But the moment someone asks for a "small change" and things start to break, it becomes clear that one step was skipped: systems analysis.

Systems analysis is what happens before anyone writes a line of code or drags a component onto a Webflow canvas. Many teams simply call it "writing the spec" or "requirements". It is the stage where you work out what the real problem is, how your systems are connected and what will happen if you change one of them. A good specification saves weeks of unnecessary development, and a missing one is one of the most common reasons projects run over budget.

In this guide we cover:

  • what systems analysis is and how it differs from business analysis;
  • which kinds of problems it helps you solve;
  • how to carry out systems analysis step by step;
  • what a good specification document should contain;
  • which methods and tools are used in 2026;
  • which mistakes are worth avoiding.

What systems analysis is

In technology, systems analysis is the work of finding and describing the best solution to a business problem while taking into account the features and limits of the systems you already have. It has no fixed deliverable: the outcome depends on the industry, the size of the business and what you were asked to solve.

Here is a familiar example. The marketing manager wants to remove one field from the contact form on the website because she believes it hurts the conversion rate. She asks the developer to delete it, and the developer does. The next day it turns out that this field was passed to the CRM, where it decided which sales rep received the lead. Now every lead lands with the same person, some of them never get a reply, and nobody understands why.

The goal of systems analysis is to predict a scenario like this in advance. You check what the field is connected to, which data it stores, where that data goes and what will happen if you remove it. Then you propose a solution that meets the marketing team's need without breaking the logic of the other systems. For example, you make the field optional and set up a default routing rule in the CRM.

The more systems talk to each other, the more systems analysis matters. A website connected to a CRM through a webhook, a CRM that sends an automatic WhatsApp message, and an invoicing tool that receives data from the payment provider are exactly where one small change snowballs into a big failure. If you want to understand how systems like these are connected technically, read our guide to REST APIs as well.

A floating glass form sends purple and light-blue light lines to several glossy modules, and one of the lines breaks off in orange sparks
One form field can affect several systems at once. Illustration: RJL Studio

Who does systems analysis: systems analysts, business analysts and everyone in between

Larger companies have a dedicated role for this: the systems analyst. This is the person who defines the requirements for a system and supports its development from start to finish. Their main tasks are to:

  • gather and analyse requirements from the business side, or investigate the problem the business came with;
  • look for ways to solve the problem or deliver the requirements;
  • write the specification, describing the requirements in a language developers understand;
  • break the work down into development tasks;
  • check that the system works correctly and meets the requirements, if the project has no dedicated QA tester;
  • write technical documentation and user guides.

Systems analyst vs business analyst

The two roles are close, but the emphasis is different. A business analyst knows the business domain in depth: how the company sells, who the customers are and how internal processes run. A systems analyst knows the technology in depth: how the system is built, how data flows and what the limits of each component are. On a good project they complement each other: the business analyst explains the "what" and the "why", and the systems analyst turns that into the "how".

What happens in a small business

Most small and mid-sized businesses do not have a systems analyst on the team, and that is fine. In practice the role is filled by whoever builds the system: a studio, a freelancer or an external developer. The important question is not their job title, but whether anyone does this work at all before building starts. If your vendor wants to start developing without asking questions about your processes and existing systems, treat that as a warning sign.

Which skills it takes

Systems analysis relies on three core skills:

  • Logical thinking: the ability to analyse information and draw conclusions from data rather than gut feeling;
  • Systems thinking: the ability to see how components are connected and understand what happens to everything else when you change one of them;
  • Communication: the ability to listen, ask the right questions and translate between business people and developers.

On top of that, you need to understand the systems involved and how software development works. A systems analyst does not have to write code, but they do have to understand what the developers are talking about.

Which problems systems analysis solves

Problems in systems analysis are usually split into three groups, depending on how much information you have about them up front.

Structured problems

These are well-known problems with plenty of data behind them. There are metrics showing that something is wrong, and you can estimate how to fix it. Say you built a booking system and wrote a manual describing every step and every error message. Based on the step where a user got stuck or the error they saw, you can tell immediately where the problem is.

Semi-structured problems

Here you have data, but it is not clear whether it is enough. Sometimes you need to collect more information, and sometimes you need qualitative signals as well as quantitative ones. Imagine the same booking system, but this time nobody wrote clear error messages. A user tries to log in and sees "Error 403". They have to figure out on their own what it means, or call support. If the code were followed by a message such as "You don't have access to this page. Please contact your account manager", the path to a solution would be much shorter.

Unstructured problems

For these problems there is almost no information. There are no metrics, and sometimes it is not even clear where to get them. These are the hardest problems, because there is no previous experience to lean on. An example: the CEO wants to see "cost per lead by branch" on a dashboard, but that number is not stored in any system, nobody knows how to calculate it and nobody knows where it should come from. Here systems analysis starts with definitions: what exactly counts as a lead, how you know which branch it belongs to and where the cost data lives.

Three floating glass cubes: one clear and orderly, one half-clouded and one full of purple fog with a light-blue beam piercing it
Structured, semi-structured and unstructured problems. Illustration: RJL Studio

How to do systems analysis: step by step

There is no single algorithm for systems analysis; everything depends on the task. But most projects go through the same stages.

Step 1. Confirm that the problem really exists

The first mistake is trusting the initial data without checking it. If the business owner says "fewer people are signing up for our newsletter", you first check GA4 or your email platform to see how many people signed up in previous months and how many sign up now. Sometimes the drop is real, and sometimes it is seasonal: around holidays and during the summer traffic changes, and that does not mean anything is broken. And sometimes it turns out that the tracking itself stopped working after a website update.

Step 2. Define the goal

If nobody set a clear task, the systems analyst defines it: which problem you are solving and what result you want to reach. A handy framework for this is SMART: a goal that is specific, measurable, achievable, relevant and time-bound. Not "increase sign-ups on the website", but "bring sign-ups back to last quarter's level within three months".

Step 3. Map the current state

Before looking for a solution, you need to understand how things work today. Which systems are involved? What information moves between them, and in which direction? Who uses each system, and why? At this stage you draw a diagram of the current process, talk to the people who work with the systems every day and check the actual settings, not just what old documents say.

Step 4. Look for a solution

This is the longest stage. The systems analyst studies the system and the relationships between its components, tests hypotheses and looks for the way to reach the goal with the fewest resources. An experienced analyst closes some questions without involving the development team at all. Suppose the sales manager asks for a new report in the CRM that does not exist yet. The analyst checks and finds that an almost identical report already exists in Google Analytics or in the ad platform. Instead of building something new, they give the manager access to the existing report, and the need is solved the same day.

Step 5. Write the specification

If the solution requires development or a change to the system, the systems analyst writes a specification for the developers. It explains the details, takes the system's connections and limits into account and identifies which parts the change will affect. We cover what such a document should contain in the next section.

Step 6. Test and document

After development you check that the solution really meets the goal defined in step 2 and that nothing broke along the way. Finally, you update the documentation: the diagram, the list of integrations and the user instructions. Without documentation, the next time someone asks for a "small change" you will have to start from scratch.

A white figure climbs a winding glass staircase with objects on the steps: a magnifying glass, a target, a map, a light bulb, a document and a glowing check mark
Systems analysis moves from a question through mapping and a solution to testing and documentation. Illustration: RJL Studio

What a good specification includes

A specification does not need to be long; it needs to be clear. A developer reading it should understand what to build without guessing. These are the sections worth including.

Goal and context

Two or three paragraphs: which problem you are solving, why it matters to the business and how you will measure success. This section helps the developer make the right call even in situations the document does not cover.

Users and permissions

Who uses the system and what each of them is allowed to do: a customer on the website, a sales rep, a manager, an accountant. Each role has different screens and actions.

Processes and scenarios

A description of every key process, step by step, including what happens when something goes wrong: a failed payment, a form submitted twice, a user who closes the browser halfway through. The easiest way to describe this is with a flowchart.

Data and fields

Which data is collected, where it is stored, what format it has and what is required versus optional. This is also where you map fields between systems: which form field goes to which CRM field.

Integrations

A list of every connection to external systems: CRM, payments, invoicing, email marketing, WhatsApp, analytics. For each connection, note the direction, the frequency, what happens on error and who owns it.

Non-functional requirements

Page speed, security, backups, accessibility and privacy. If you work with customers in the EU, this usually means designing for the GDPR from the start: deciding which personal data you really need, where it is stored, who can access it and how long you keep it. Email and tracking consent also fall under ePrivacy rules, and many digital products and online shops now need to meet accessibility requirements under the European Accessibility Act, with WCAG as the usual reference. This is a general overview, not legal advice, so check the details with a qualified adviser for your market.

Acceptance criteria

How you will know the work is done. A list of defined checks, for example: "A lead from the homepage form appears in the CRM within one minute, with the traffic source and campaign name". Criteria like these save a lot of arguments at the end of a project.

A floating glass blueprint document with abstract rounded lines and glowing check marks, surrounded by 3D icons of a user, a padlock, a gear and a plug
A good specification answers the developer's questions before they ask them. Illustration: RJL Studio

Systems analysis methods and tools in 2026

There are many methods and tools, and analysts choose between them based on the task and their own experience. These are the most widely used.

Business process modelling: BPMN

The core method in systems analysis is modelling business processes. The most common standard is BPMN (Business Process Model and Notation): a set of agreed symbols that lets you describe a process as a diagram, with tasks, decisions, events and who is responsible for each step. Alongside BPMN, teams also use simple flowcharts, which are enough for most small-business projects. We wrote about managing and improving processes in our guide to business process management.

UML and data diagrams

UML (Unified Modeling Language) is a modelling language that describes the structure of a system: which components it has and how they interact. The most common diagrams are component diagrams, sequence diagrams (who sends what to whom, and in which order) and state diagrams. To describe the data structure, analysts use an ERD (entity relationship diagram), which shows which tables or collections exist and how they relate to each other. This is especially useful when you plan a CMS with many collections and references, for example in Webflow or WordPress.

Diagramming and collaboration tools

Any software that lets you draw processes and connections works for systems analysis. The most popular options in 2026:

  • Miro and FigJam: collaborative boards for client workshops and quick diagrams;
  • draw.io (diagrams.net) and Lucidchart: clean flowcharts, BPMN and UML diagrams;
  • Camunda Modeler: professional BPMN modelling, including processes that run automatically;
  • Notion and Confluence: writing specifications and documentation;
  • Jira, Linear or monday: managing the development tasks that come out of the specification.

Data tools

When the analyst works more with data than with diagrams, they use a different toolkit: SQL for querying databases, Google Sheets for quick checks, and Looker Studio or Kibana for visualising data and logs.

What about AI

In 2026, tools such as ChatGPT, Claude and Gemini help you draft a first version of a specification, turn a meeting transcript into a list of requirements and suggest edge cases you forgot. That saves time, but it does not replace the work itself: the model does not know your systems or the decisions made in the past. Everything it writes has to be reviewed by someone who knows the business. And do not paste real customer data into these tools.

A floating 3D flowchart made of rounded glass shapes, decision diamonds and circles, connected by glowing purple and light-blue arrows with orange spheres moving along them
A process diagram is the fastest way to explain how a system works. Illustration: RJL Studio

Systems analysis for websites and automations: real-world examples

Systems analysis sounds like something for big tech companies, but it matters just as much for a small business website. A few common situations:

A lead form connected to a CRM

A Webflow or WordPress site sends leads to HubSpot or monday CRM, which then triggers an automatic WhatsApp message to the customer and the sales rep. Before you change anything in the form, you need to know which fields are used for routing, which for reporting and which for automations. A one-page diagram of every field and its destination is enough to prevent most failures.

An online shop with payments and invoicing

An order on the website triggers a payment with the payment provider, an invoice, a stock update and a confirmation email. Any of these steps can fail. The specification should define what happens if the payment went through but the invoice was not created, or if the stock level did not update. Without that, you only find out about problems when a customer complains.

Automations in n8n, Make or Zapier

Automations are quick to build, which makes it easy to build them without thinking. A year later you have twenty scenarios and nobody remembers who built them or why. Systems analysis here means documenting every automation: what triggers it, what it does, which systems it touches and who gets an alert when it fails. We wrote about building automations like these in our n8n guide.

Moving to a new website

When you migrate from one platform to another, for example from Wix to Webflow, moving the content is not enough. You need to map everything the old site quietly did: forms, tracking pixels, integrations, 301 redirects and pages that bring in search traffic. Whatever is not mapped gets lost.

Common mistakes in systems analysis and specifications

Starting with the solution instead of the problem

"We need an app" or "add a chatbot" are solutions, not problems. Before writing a spec, ask what you are trying to achieve. Sometimes the problem can be solved with a small change to the existing website.

Trusting what is written instead of what actually happens

Old documents, settings someone remembers by heart and "it has always worked this way". Check the actual settings in the systems themselves.

Forgetting the edge cases

The happy path, where everything works, is the easy part. Most bugs come from cases nobody described: an empty field, a double payment, a customer who comes back a week later.

Not involving the people who use the system

The sales reps, office staff and team leads who use the system every day know things no senior manager knows. Half an hour talking to them saves weeks of fixes.

A spec written once and then forgotten

Systems change all the time. If the specification and documentation are not updated after every change, they become irrelevant, and next time you start from zero.

How RJL Studio specs and builds systems for businesses

At RJL Studio every project starts with a short call and a mapping session: what the business needs to achieve, which systems are already in place and how information flows between them. When it is needed, we prepare wireframes and process diagrams before building starts, so everyone agrees on the same picture. Only then do we build the website in Webflow, WordPress or Wix, connect it to your CRM, payments and other systems through APIs, webhooks and n8n, Make or Zapier, test everything and document it.

Want to connect your business systems without breaking what already works? Read about our automation and integration services and UX/UI design, or get in touch and we will get back to you.

Frequently asked questions

What is the difference between systems analysis and a specification?

In practice the two terms are often used almost interchangeably. Strictly speaking, systems analysis is the process: investigating the problem, mapping the systems and looking for a solution. The specification is usually the output: the document that describes exactly what needs to be built. There is also a distinction between a business specification, which focuses on needs and processes, and a technical specification, which focuses on how the system is structured.

Does a small business need systems analysis?

Yes, at a scale that fits it. A small business does not need a fifty-page document, but it is worth having one diagram of your systems and the connections between them, plus a list of your automations and who owns each one. That is enough to prevent most failures when you change something.

How long does a specification take?

It depends on the scope. For a change to an existing integration, a few hours can be enough. For a company website with a form and a CRM connection, it usually takes a few days. For a system with a customer portal, payments and many integrations, the specification can take several weeks. Either way, that time is saved later during development.

Can you write a specification with AI?

You can use AI to write a draft, structure requirements from a meeting transcript or suggest edge cases. But the tool does not know your systems or your processes, so everything it writes needs to be reviewed by someone who knows the business. Do not feed customers' personal data into tools like these.

Who should write the specification: the client or the vendor?

Ideally both, together. The client knows the business, the customers and the processes, and the vendor knows the technology and its limits. A good vendor will ask a lot of questions, propose solutions and write the document, and the client will review and approve it before building starts.

Disclaimer: blog content is for general information and learning only and is accurate as of its publication date. It is not professional, legal, financial or business advice, nor a recommendation or call to take any action. Any use of this information is at the reader's own risk, and we accept no liability for any resulting loss. Links to external websites and tools are provided for convenience only. We recommend consulting a qualified professional before making decisions.