What marketing can learn from agile attribution and software engineering

Marketers can learn from software engineering about how we think about using technology to improve performance. In 2001, a team of thought leaders in software engineering analysed project failures in the industry and re-conceptualized how to deliver software.

Part of the problem was the commonly used ‘Waterfall methodology’.

  1.  Analyse (figure out the problem)
  2.  Design (figure out the blueprint to solve the problem)
  3.  Build (execute on the blueprint)
  4.  QA (check you’ve solved the problem successfully)

It seems logical. But the problem with such a structured methodology was that it provided little flexibility for change. Major software projects to solve complex problems, often took such a long time to complete that once finished, the environment (market or technology) had moved on, meaning the entire project was either obsolete or required major ‘re-factoring’.

Alternatively, business stakeholders involved with sponsoring the project may no longer be around, making it difficult to assess whether the results were successful or still strategically relevant. If problems arose within the QA process, important domain knowledge about the problem was often no longer available.

A 2013 study from Ambysoft found that less than half (49%) of projects following a traditional waterfall structure were successful.

As a result of these shortcomings, the ‘Agile’ approach was born. Almost two decades later it’s clear there are lessons that Marketers can take from Agile’s start small, iterate, and learn approach to make sure effort and cash is expended in the most ROI positive way possible.

Unlike waterfall models, an agile approach allows businesses to derive learning and ROI before a product is ‘finished’.

The Agile approach focuses on building software in small chunks, with the emphasis on delivering working software in ‘sprints’ which would take a few weeks. At the end of each sprint, the team pause to review their objectives with their customers or stakeholders and course correct as necessary. QA is carried out in parallel with progress, allowing basic products – known as ‘minimal viable product’ – to be deployed within a much shorter timescale, then tweaked and iterated updated based on actual customer usage and feedback.

How is this applicable to marketing?

One of the constants of every industry is change. The fundamental problem of waterfall is that the ‘requirements’ of the software are liable to change during the design and build phases, leaving the end-product effectively solving yesterday’s problems. The inability to test the product before it has passed QA also means that none of its features can be used or benefited from during the entire project.

This pattern has clear parallels in marketing measurement technology, where marketers invest budget into expensive platforms that can take many months to become useful. We should then, look at applying the principles of Agile software engineering.

Individuals and interactions over processes and tools

We observe an anti-pattern in tech buying which we call ‘Can Do vs Does Do’. In these cases, marketers have been sold expensive technology platforms that salespeople claim ‘can do’ everything they need. What they don’t mention (or don’t understand themselves) is the necessity of humans to make sure the product ‘does do’ what it needs to. For Marketing Technology companies, people present additional cost and business model complexity compared with the high profit and low cost of deploying and maintaining SaaS (software as a service) products.

However, given the complexity of the marketing channel environment, and the constant state of disruptive technological and regulatory change, the notion that a self-service SaaS tool can be all things to all people is self-evidently ridiculous.

From helping to set up data source integrations, to a friendly voice on the end of the phone, managing, guiding and advising, you need people alongside technology to make sure it ‘does do’ what you need. These people need to understand your problems and objectives deeply – it’s about individuals and interactions over processes and tools.

Working software over comprehensive documentation

Waterfall was obsessed with ‘requirements gathering’ (documenting everything the software needed to do) before building begins. This means that by the time the requirements are ‘gathered’, built, and tested they may no longer be required. A parallel in marketing is ‘360-degree attribution’ where marketers spend time and money trying to integrate and model all online/offline channels before taking actions to improve.

Whilst it is sensible and desirable to build towards a complete view of marketing attribution, it is equally important to avoid ‘analysis paralysis’ – addressing your problems in an Agile way that produces tangible return on investment within a few months.

Customer collaboration over contract negotiation and responding to change over following a plan

At the beginning of an engagement with a marketing attribution vendor, it’s a good idea to create a roadmap which addresses your objectives to remove any ambiguity from the relationship. This is crucial, to describe and write down a shared understanding of what you need, what the vendor will do, the specific format of delivery, and the key result required.

However, an effective relationship must be able to respond to changes in priorities and the external environment. There are constants in your objectives which won’t change (you will always want to improve customer acquisition cost and increase return on advertising spend) but the steps taken to achieve these goals must be flexible. For example, something that was possible 12 months ago may not be possible now (see removal of access to Doubleclick user IDs, or user level Facebook impression data), or a vendor might have new answers and better ways to help you (such as  algorithmic multi-touch attribution modelling or using AI to build audience clusters).

To adapt to unforeseen challenges and leverage technology and data science, a collaborative approach is required with your tech partners, with a roadmap that adapts as your needs evolve. At Fospha we call this ‘Agile and accessible attribution’ – its about finding solutions together and quickly transform analysis into action for your brand.