Before we start, what do I mean by Optimisation (with a capital O)?
I think of Optimisation as a structured data-driven approach to improving the digital customer experience. Covering both testing (A/B or multivariate) and personalisation – as long as you’re trying new things and measuring their impact, that’s Optimisation.
Almost all business go about optimisation (lower-case o), but in an unstructured way. Once you formalise your business processes around testing and personalisation, you can maximise your impact through knowledge management, operational efficiency and scaling.
There are two prevailing Optimisation methodologies:
Crawl, Walk, Run
Big Bang Use Cases
To use a ham-fisted analogy, imagine you’re building a wall. In Crawl-Walk-Run, you use many small bricks (experiments), and in Big Bang Use Cases, you use fewer large blocks (experiences).
So which is better?
Crawl-Walk-Run (call it Walk-Run-Fly to your stakeholders if you want to sound faster) is an iterative methodology for growing your Optimisation capability and impact.
In this methodology, you drive your progression through evolving your team’s maturity. Start small, then take on larger and more complex use cases as you grow your capability and refine your processes.
Starting in the Crawl phase, your objective is to execute a few quick wins and begin to develop a playbook for ongoing optimisation.
I recommend drafting a one-pager on how you’ll being to approach optimisation (who is involved, scope and broad responsibilities) but don’t let documentation and process design come before results. You want to be trying, succeeding and failing quickly – it’s the best way to learn what works for your business and your customers, and your learnings here are what step you into the Walk phase.
Throughout all phases, you’ll want to run regular ideation and prioritisation sessions. Engage a broad range of stakeholders and get as many ideas for testing and personalisation as you can, then plot against impact and effort (don’t worry if you’re guessing, this will get easier over time as you dial in your estimations.
You’re looking for low effort experiments to execute, learn from and move on.
e.g. why not start with the quintessential A/B test – do we get more conversions if our CTA button is green instead of red? – probably not, but you won’t know until you try it.
In the Walk phase, you’ll build from your foundations, use insights to increase your success rate, and formalise your Optimisation processes.
By now, you’ve run at least 10 experiments. Keep the winners going and kick the losers to the curb (after logging your results).
The key to the walk phase is refining your Optimisation business processes. You’ll keep developing and releasing experiments throughout, but now you can use your experience to streamline your stakeholder engagement, prioritisation, experimentation, and reporting processes to scale your impact.
And remember, knowledge is the key to progression. Document all this knowledge and share it far and wide – maturity is business-wide not individual.
In the Run phase, you’ve hit max velocity. You’ve got your processes streamlined, you’re hitting more winners than losers, and you’ve got buy-in and advocacy across the organisation.
Now is the time to take on larger and more complex experiments, drawing from your high impact/high effort quadrant. And if you haven’t already, you should measure the impact of your Optimisation program as a whole on your organisation’s bottom line.
Benefits of Walk-Crawl Run
Shorter time to first realised value – Small experiments during the Walk phase will begin delivering business value quickly.
Self-funding – Early experiments often deliver enough revenue for the Optimisation program to fund itself.
Incremental buy-in – Starting small requires low level of buy-in from other teams. Then once you ramp up, and can show value, it’s easier to gain buy-in.
Iterative insights – The more experiments you do, the more you learn and can apply to future experiments.
Minimal investment – Only a small team may be needed in the crawl phase, requiring minimal investment and the ability to run in a BAU team.
Drawbacks of Walk-Crawl-Run
Not suited to large use cases – In an iterative Optimisation cadence, you may reach a ceiling when your current customer experience is streamlined and effective, and improve further you’d need to redesign the fundamental experience.
Big Bang Use Cases
Where Crawl-Walk-Run methodology focuses on small-scale iterative Optimisation, Big Bang Use Cases take a large-scale transformative approach to optimised customer experiences.
In this methodology, you’ll design a new or existing customer experience with personalisation in mind. More than UI changes, you’re most likely taking a major customer pain point and changing the way that they interact with your business.
e.g. you might rebuild your e-commerce experience with personalised product and content recommendations to prompt cross-sell and engagement.
Each Big Bang Use Case has the potential to drive significant business value, but the effort and complexity make them a major undertaking. So this suits a project approach with high organisational buy-in and governance.
Analysis is key here; you need to be confident that your selected use case will have a concrete impact on your business’ bottom line.
Benefits of Big Bang Use Cases
High impact – transformative customer experiences have the potential to drive greater impact than many smaller optimisations.
Visible optimisations – with large changes, regular customers will notice and appreciate when their experience changes for the better.
Building foundations for additional optimisations – enabling your use case will usually involve building technology foundations that can be leveraged to additional use cases.
Drawbacks of Big Bang Use Cases
Risk of failure – despite your initial analysis, success is not guaranteed. If your completed use case doesn’t increase conversion, you may be left with a high regret spend.
Whichever methodology you choose, remember to set clear business objectives before starting your Optimisation program. And measure yourself against those objectives regularly.
Set up your data foundations early to enable initial and future use cases. The impact of your personalisation will always be limited by the accuracy and breadth of your customer data. So whether you’re using a data layer, Customer Data Platform or other data source, you’ll need to maintain good data management practices and prepare for extensibility and future use cases.
Similarly, consider the capabilities and limitations of your technology enablers – if you have or are acquiring a personalisation engine, like Salesforce Marketing Cloud Personalization, how will it fit into the rest of your technology stack and how will you use it to develop value?
And finally, know that these two methodologies are not mutually exclusive, although it will take extra resources to manage large experiences alongside small experiments.