Catch-22 within test automation – the dangers of surrogation

25-minute Talk

In testing and test automation, we sometimes focus too much on creating tests and lose sight of our true goal. This is the danger of the surrogation effect and supernormal stimuli.

Virtual Pass session

Timetable

11:45 a.m. – 12:30 p.m. Friday 22nd

Room

Room F3 - Track 3: Talks

Audience

All software professionals

Key-Learning

  • Sometimes goals can be substituted via a process of surrogation.
  • The surrogation effect is driven by having inappropriate measures.
  • Keep your goal valid, by addressing the organisational and measurement issues that drive teams towards surrogation.

Why, in the novel Catch-22, did the American warplanes bomb a shoal of fish in the sea, rather than the enemy target assigned to them? And why did their commander celebrate this as a measure of his squadron’s bombing skills, rather than criticise their flagrant waste of Allied resources? And what on earth does any of this have to do with testing and test automation?

These things are all connected by the phenomenon of surrogation.

Organisations invest in test automation with a clear goal in mind, then develop metrics to measure progress towards that goal. But sometimes an automation team loses sight of this original goal and instead substitutes their own surrogate goal of maximising their score on the metrics.

In this session, we start by looking at why the characters within Catch-22 abandoned their assigned goal and instead drifted into the bizarre behaviour of bombing fish. We then use this knowledge to understand the problems we encounter when automation teams lose sight of their original goal and substitute instead a goal of maximising the measures that were originally intended to drive their progress towards that goal.

We explore what drives teams into this dysfunctional behaviour, plus its consequences. We look at attribute substitution, the mechanism that allows surrogation to occur. We then identify the conditions typically present within an organisation that increase the risk of surrogation occurring. We also look at how our choice of metrics, team structure, plus other factors, like absence of psychological safety, can unintentionally drive behaviour towards surrogation. Next, we look at the possible consequences of surrogation upon the true value of the automation suites that we build.

We conclude by showing how organisations can address this problem, through restructuring their development teams, plus constructing their metrics and goals in ways to minimise the risk of automation for its own sake becoming an unintended goal.

Come along to our session and learn how to avoid your test team bombing fish in the sea!

Related Sessions

Virtual Pass session
11:45 a.m. – 12:30 p.m.
Room F3 - Track 3: Talks

25-minute Talk

2:45 p.m. – 4:45 p.m.
Room D3+D4 - Track 8: Workshops

120-minute Workshop

Virtual Pass session
5:15 p.m. – 6:15 p.m.
Room F1+F2+F3 - Plenary

45-minute Keynote

Virtual Pass session
4:00 p.m. – 4:45 p.m.
Room F2 - Track 2: Talks

25-minute Talk