Skip to content
Tutorials Center
My learning

Topics

Lesson 7 of 50 ยท The SDLC and Process Models

Spiral Model and Risk-Driven Development

Learn how the spiral model runs repeated risk analysis loops, uses prototypes to reduce uncertainty, and weighs the cost of reducing a risk against its benefit.

Every project has things that could go wrong: the requirements may be misunderstood, a technology may not scale, a key person may leave. Most process models mention risk in passing. The spiral model, proposed by Barry Boehm in the 1980s, puts risk at the center. The team works in loops, and in every loop it asks the same question first: "What is the biggest threat to this project right now, and what is the cheapest way to reduce it?"

The spiral model is a way of organizing decisions, so there is nothing to install or run. The examples below are plain JavaScript stand-ins with invented numbers. They show how risk exposure and risk-reduction decisions are calculated; they are not a real project-management tool.

The idea in one picture

Imagine a spiral drawn on a page. The distance from the center is the money and effort spent so far. Each full turn around the center is one loop of work. Early loops are small and cheap and aim to learn things; later loops are bigger and aim to build and deliver.

Every loop passes through four activities, which are usually drawn as the four quadrants of the spiral:

  1. Set objectives. What does this loop have to achieve? What are the constraints (budget, deadline, technology) and the alternatives?
  2. Identify and resolve risks. List the uncertainties, rank them, and take action on the worst ones: build a prototype, run a simulation, interview users, benchmark a database.
  3. Develop and verify. Build the piece of the product (or the prototype) that this loop calls for, and check it.
  4. Plan the next loop. Review the results with the stakeholders, decide whether to continue, and commit to the next loop.
flowchart LR
    A["1. Set objectives<br/>and alternatives"] --> B["2. Identify and<br/>resolve risks"]
    B --> C["3. Develop<br/>and verify"]
    C --> D["4. Review and<br/>plan next loop"]
    D -->|"next loop, bigger scope"| A
    D -->|"risk too high or value gone"| E["Stop or change course"]

One loop of the spiral. The arrow back to step 1 repeats with a larger scope; the exit arrow is the important safety valve.

Four steps in every loop

Objectives What must this loop achieve?
Risks Rank them, reduce the worst.
Build Make and check the piece.
Plan Decide whether to continue.

The loop repeats, and each turn costs more than the last.

1 / 2

The review in step 4 matters as much as the building. A project may legitimately end after loop two, because a prototype showed that the idea is technically impossible or that nobody wants it. Stopping at that point costs a small fraction of what a failed waterfall project costs at the end. A process that makes stopping easy is protecting the budget.

A walk through four loops

Here is how a spiral might look for a mobile app that routes delivery drivers. The details are invented, but the pattern is typical.

Loop Main risk Action Outcome
1 Do drivers want it at all? Interview ten drivers, build clickable screens Drivers want turn-by-turn order, not a map
2 Can routing run fast enough on a phone? Prototype the algorithm and benchmark it Fine for 30 stops, too slow for 300
3 Can the backend handle 2,000 drivers at once? Load-test a thin slice of the server Needs a queue; design changes
4 Delivery, support, polish Build the first release incrementally Released with the risks retired

Notice that the risks are different in each loop, and that the early loops produced knowledge rather than product. By loop 4 the project resembles an ordinary iterative and incremental build, because the dangerous unknowns are gone. That is the typical shape: the spiral front-loads the learning and then converges on normal development.

Measuring risk: probability times impact

To rank risks you need a number. The classic measure is risk exposure:

TEXT
risk exposure = probability of the loss x size of the loss

If there is a 30% chance that the routing algorithm is too slow, and fixing that late would cost 40 person-weeks, the exposure is 0.3 x 40 = 12 person-weeks. This is an expected loss, a statistical average rather than a prediction for this one project, but it lets you compare very different risks on one scale. The risk at the top of the list gets attention first. You will see the full process in risk management.

This stand-in ranks a small risk list and prints the exposure of each one.

JS Rank risks by exposure

Press Run (or Ctrl + Enter) to see the console output here.

No output.

The ranking is a conversation starter, not an oracle. The probabilities are guesses; what the table does is force the team to state its guesses and argue about them in the open.

Is reducing a risk worth it?

Reducing a risk is not free. A prototype costs time, a load test costs infrastructure, and an expert review costs money. The spiral model asks for a cost-benefit comparison:

TEXT
risk reduction leverage = (exposure before - exposure after) / cost of the action

A leverage above 1 means that the action is expected to save more than it costs. The higher the number, the better the deal. A leverage below 1 means the action costs more than the risk it removes, so you should accept the risk (and maybe watch it) instead.

Leverage is only as good as the estimates behind it. When two actions have similar leverage, prefer the one that also teaches you something, because information is useful even when the original guess about the risk was wrong.

The next stand-in compares four possible actions. Each lowers the probability of its risk, and each has a cost.

JS Risk reduction leverage

Press Run (or Ctrl + Enter) to see the console output here.

No output.

The benchmark has the highest leverage because it is cheap and removes most of the exposure. The retention bonus costs more than it saves in expectation, so the model says to accept that risk. A team might still decide to pay it for non-numeric reasons, and that is fine, as long as the decision is made on purpose.

Prototypes in the spiral

A prototype is the spiral's favorite tool, because it turns an uncertain guess into evidence. The loop table above used three kinds:

  • A user-interface prototype (clickable screens) reduces requirements risk.
  • A technical prototype (a spike that benchmarks one risky component) reduces technology risk.
  • A vertical slice (one thin feature through every layer) reduces integration and architecture risk.

Decide beforehand which question the prototype must answer and what result would change your decision. A prototype built without a question tends to grow into the product by accident (see technical debt).

Strengths and weaknesses

Strengths: it handles uncertainty explicitly, it gives early exits, it works well for large, expensive or novel projects, and it mixes naturally with other models inside a loop (an iteration can be agile).

Weaknesses: it needs people who can recognize and estimate risks, which is a real skill; the repeated analysis and reviews add overhead that is hard to justify on a small, well-understood project; and contracts with fixed prices fit it poorly because the scope of later loops is not known up front. It can also degrade into endless loops of analysis if nobody has the authority to say "enough, build it".

The spiral model is rarely followed by the book today, but its central habit survives everywhere: do the riskiest, least-understood work first, and reduce uncertainty before spending heavily. Modern teams call these "spikes", "tracer bullets" or "walking skeletons". If you cannot yet decide which model fits a project, the next lesson gives a method.

Check your understanding

Quick check ยท 5 questions

of 5 answered You scored / 5 ยท perfect!

  1. 1. What is the defining feature of the spiral model?

    The spiral model is risk-driven: each loop sets objectives, resolves the largest risks, builds, and then reviews and plans the next loop.
  2. 2. A risk has a 25% probability and a loss of 40 person-weeks. What is its exposure?

    Exposure is probability times loss: 0.25 x 40 = 10 person-weeks.
  3. 3. An action costs 4 person-weeks and lowers a risk's exposure from 12 to 6. What does its leverage say?

    Leverage = (12 - 6) / 4 = 1.5. Above 1 means the expected saving is larger than the cost.
  4. 4. Why is the review step at the end of every loop important?

    Early loops are cheap, so a bad idea can be dropped at a fraction of the cost of finding out at the end.
  5. 5. When is the spiral model a poor fit?

    Small, well-understood projects do not repay the overhead of repeated risk analysis, and fixed-price contracts do not fit open-ended later loops.

Summary

  • The spiral model repeats four activities in every loop: set objectives, identify and resolve risks, develop and verify, review and plan.
  • Early loops buy knowledge with prototypes and spikes; later loops behave like ordinary iterative development.
  • Risk exposure is probability times loss; reduction leverage compares the exposure removed with the cost of the action.
  • Every loop is an exit point, which keeps the cost of a doomed project small.
  • It suits large, novel, uncertain projects and is overkill for small, well-understood ones.

Next: Choosing a Process Model.