DXPlaybook
Play 6: Measurement & Optimisation
DXPlaybook is Codehouse’s practical guide to running an enterprise‑grade digital experience with less drama and more certainty. It is written for leaders and senior specialists across marketing, product, digital, content and design, with enough depth that delivery teams can act on it. Each play turns a fuzzy ambition into something visible, ownable and repeatable.
This page is Play 6. It focuses on measurement and optimisation — how decisions are made, what data drives improvement, and how to build a culture of experimentation and continuous learning. The scope spans analytics, user research, controlled tests and personalisation without overcomplicating things. Reliable measurement links effort to outcome. Harnessing analytics to understand visitor behaviour and campaign performance is essential for creating meaningful digital experiences. When you measure what matters and act on those insights, you can achieve significant gains in engagement and targeted interactions. Play 6 gives you the principles and questions you need to turn data into decisions and iterate with confidence across channels.
Why this play matters
Play 6 is about more than just collecting data. It’s about creating a culture where decisions are driven by solid evidence and continuous learning. In many B2B organisations, data collection happens in silos—marketing looks at campaign metrics, product teams monitor usage, content teams track engagement—but rarely does anyone connect the dots to understand how all these signals tie back to business goals. When metrics are shallow or inconsistent, people default to assumptions, launches drag on, and optimisation becomes guesswork.
Measurement matters because it links effort to outcome. Good analytics and optimisation practices tell you not only what is happening on your site or app, but why. For example, Codehouse’s work with the Endeavour Foundation leveraged Sitecore’s analytics and personalisation capabilities to build a deep understanding of diverse audiences, capture user behaviour and use A/B testing to tailor experiences. This approach delivered a 31 % increase in visitor engagement value and a 151 % uplift in targeted interactions—results that wouldn’t have been possible without reliable data and an optimisation mindset. That case shows how measuring the right things and iterating deliberately can transform both user satisfaction and business value.
Finally, robust measurement practices are the foundation of experimentation, personalisation and predictive analysis. They help teams explore what works through controlled tests, and they build confidence in scaling what works across channels. Without trustworthy data and clear success metrics, experimentation stalls, personalisation misfires, and budgets are wasted. Play 6 lays out the principles and questions leaders should use to ensure that analytics, experimentation and optimisation are serving the overall vision, not working at cross purposes.
Questions to ask
This section helps leaders diagnose how measurement and optimisation support decision‑making and continuous improvement. It is written for marketing, product and digital leaders who need to act on insights without getting lost in data jargon. Use it as a workshop checklist or work through it asynchronously. For any issues you plan to act on, capture Owner, Evidence link, Status and Next step so improvements are visible and accountable.
Quick triage
These high‑level questions reveal whether measurement and optimisation are working. Address them before diving into detail.
What decisions are currently made based on data, and where do teams still rely on gut feel or assumptions?
Do we have a single source of truth for digital metrics and agreed definitions across teams?
How quickly do we measure the impact of a change or campaign — in hours, days or weeks — and is that fast enough to inform decisions?
Do stakeholders trust the numbers enough to change course when the data suggests it?
Measurement landscape and ownership
Understanding what tools are in place and who owns them is the first step to improving them.
Which analytics, experimentation, survey and reporting tools are in use, and who owns each platform day to day and at an accountable level?
Where are there gaps or overlaps in responsibility that lead to duplicated instrumentation or blind spots?
How do teams instrument new interactions, forms and events, and is there a standard data layer that feeds all tools?
Are there clear service‑level agreements for data availability and freshness, and do we monitor them?
Goals, KPIs and definitions
Clear goals and agreed metrics align teams and link digital effort to business outcomes.
What business objectives are we trying to move — acquisition, engagement, retention, revenue — and what digital KPIs map to each?
Do we distinguish between leading indicators (for example, engagement scores) and lagging indicators (for example, conversions or renewals), and who signs off on them?
Are KPIs consistent across marketing, product and digital teams, or does each function measure success differently?
How often do we review KPIs and refine them as user behaviour and organisational goals evolve?
Data quality, privacy and governance
Reliable data is the foundation of evidence‑based decisions. Poor quality erodes trust and invites legal risk.
How do we monitor data quality, validate at capture and reconcile across systems to avoid duplicates or gaps?
Are consent signals and privacy choices applied consistently at the top of the stack, and do we block non‑essential tags until consent is present?
Where do we store personally identifiable information, how long is it retained, and are we compliant with regional regulations?
Do we have documented retention and deletion policies, and are they enforced across all platforms?
Experimentation, testing and optimisation
Experimentation makes learning systematic. This lens reveals whether you have the tools and culture to test and improve.
Do we run controlled experiments — A/B or multi‑variant tests — to answer questions about user behaviour and content, and do we measure them properly?
How are hypotheses formulated, results evaluated and decisions made following an experiment?
Do we maintain a backlog of optimisation ideas, and is there a prioritisation framework to decide what to test next?
What prevents us from experimenting more (for example, tool capability, traffic volume, cultural barriers or regulatory concerns)?
Decision‑making and culture
Measurement only matters if it changes behaviour. These questions probe how data flows into decisions and who owns that process.
Who is responsible for synthesising insights from analytics, research and experiments, and how do they communicate those insights to decision‑makers?
Does the organisation encourage experimentation and accept that small failures lead to larger learnings, or does fear of failure suppress testing?
How do we share lessons learned across teams so knowledge compounds rather than being siloed?
When the data points in a new direction, do leaders adjust plans and budgets accordingly, or does anecdote override evidence?
Red flags to watch
Decisions are made on instinct because data is incomplete, conflicting or untrusted.
KPIs are misaligned or unknown across teams; success means different things to different functions.
There is no documented data layer or standard instrumentation; events and attributes vary by page or tool.
Consent is handled inside a downstream tool rather than via a central consent management platform and tag manager.
Experiments are rare, have no clear hypothesis or result evaluation, or are ignored when they contradict expectations.
There is no owner for measurement and optimisation, so issues linger and improvements lack accountability.
Good patterns
The following patterns help organisations measure performance reliably and turn insights into action. They apply to enterprises of all sizes and avoid locking you into specific tools. Adopt them at the pace your culture can absorb; measurement maturity grows through practice, not policy.
Make metrics meaningful
Measurements are only useful if they reflect the outcomes you care about. Align everyone around a shared vocabulary and make sure each metric answers a real business question.
Define a small set of leading and lagging indicators for acquisition, engagement, retention and revenue, and link them to your broader business objectives.
Publish a measurement framework that explains what each KPI means, how it is calculated and who owns it; review it regularly to ensure relevance.
Distinguish between metrics that signal user intent (for example, engagement value) and those that confirm success (for example, conversions), and make both visible.
Avoid vanity metrics; focus on measures that teams can influence and that correlate with customer value.
Instrument once, use everywhere
Consistent instrumentation means you capture an event once and send it to every system that needs it. A documented data layer reduces duplication and confusion.
Maintain a clear data layer schema with event names, parameters and definitions; update it when new features or campaigns require new events.
Collect consent at the top of the stack and include consent state and context with every event so analytics, marketing and personalisation respect user choices.
Capture events for view, interaction, error and success across journeys; feed analytics, experimentation and marketing automation from the same stream.
Avoid hard‑coding tracking logic in individual pages or tools; abstract instrumentation so changes propagate automatically.
Build a single source of truth
A shared view of performance prevents conflicting reports and speeds decision‑making. Bringing data together makes patterns visible.
Integrate analytics, marketing automation and customer relationship management data so you can trace a journey from first touch to revenue.
Centralise reporting in dashboards that pull from the same underlying data; give teams self‑service access to explore their own questions.
Set service‑level agreements for data freshness and monitor them; stale or missing data undermines confidence in decisions.
Keep a change log for metrics and dashboards so everyone understands when definitions change and why.
Test systematically
Experiments reveal what works and what doesn’t. Make testing a habit rather than a one‑off project.
Maintain a backlog of hypotheses with expected impact, required sample size and priority; prioritise tests that can inform high‑value decisions.
Run controlled experiments (A/B or multi‑variant) with clear success criteria and a minimal viable audience; analyse results thoroughly and document learnings.
Apply the same rigour to personalisation; treat personalisation rules as experiments and measure uplift before rolling out widely.
Share outcomes, including failures, so learnings compound across teams; avoid repeating tests that have already been run elsewhere.
Act on insights, not just reports
Reporting is where measurement starts, not where it ends. Tie insights to action and make improvement loops explicit.
Schedule regular reviews where teams interpret metrics, identify drivers and decide what to change; include both quantitative and qualitative data.
Pair dashboards with narrative summaries that explain what changed and why; avoid letting dashboards become passive “wallpaper.”
Turn insights into tickets or backlog items with owners and due dates; track whether actions based on data actually happen.
Celebrate improvements and treat negative results as learning opportunities; cultivate a culture that values evidence over instinct.
Respect privacy and consent
Data collection and experimentation must respect user choices and legal obligations. Privacy compliance is a non‑negotiable constraint, not an afterthought.
Implement a consent management platform at the top of the stack and enforce consent via your tag manager across all analytics, marketing and personalisation tags.
Anonymise or pseudonymise personal data wherever possible, and enforce data retention limits aligned with regulations and user expectations.
Audit data flows regularly to ensure that events do not capture personally identifiable information inadvertently; fix breaches promptly.
Include consent and privacy status as part of your metrics so you can see the impact of opt‑in and opt‑out rates on performance.
Build capability and culture
Measurement is a team sport. Invest in the skills and practices that make optimisation sustainable.
Train teams in analytics fundamentals, experimentation design and interpretation; make sure everyone understands the data layer and measurement frameworks.
Hire or develop analysts who can bridge marketing, product and technology and who can translate insights into action.
Encourage cross‑functional collaboration on measurement projects; pairing analysts with marketers and product owners yields richer insights.
Normalise experimentation and reporting rhythms; make them part of the delivery cadence rather than an add‑on.
Keep measurement lightweight—start with the essentials, avoid over‑instrumentation and revisit what you track periodically as priorities evolve.
Adopting these patterns will help you shift from ad‑hoc reporting to a reliable, evidence‑based practice. They work whether you use off‑the‑shelf analytics tools, a customised data platform or a composable stack. The goal is to make measurement an enabler of better customer experiences and business outcomes, not an isolated report.
Case study
The Endeavour Foundation, a long‑established Australian organisation supporting people with disabilities, wanted to revitalise its digital presence and create more personalised experiences. Its goals were to deepen understanding of diverse audiences, employ a user‑centred personalisation strategy and harness analytics to guide improvements. After a competitive selection process, a digital partner was appointed to lead the transformation.
The project began with an optimisation blueprint. Collaborative work mapped customer journeys and designed a content strategy centred on personalisation. A suite of personalisation and testing tools was implemented to target segments and run controlled experiments. Analytics captured engagement and conversion data, allowing the team to measure which experiences resonated and adjust tactics quickly.
The results were striking: the initial rollout delivered a 31 % increase in visitor engagement value and a 151 % uplift in interactions with personalised content. Personalised banners and pathways guided visitors to high‑value pages, and real‑time behavioural data fed back into the optimisation process. Knowledge transfer was a priority, enabling the organisation’s own teams to continue iterating independently.
This example shows why measurement and optimisation must go hand in hand. By setting clear goals, using analytics to understand user behaviour and testing variations systematically, the programme proved that targeted digital experiences can drive significant gains in engagement and value. It demonstrates how a disciplined approach to measurement validates personalisation efforts and inspires further innovation.
Signals and maturity
This section makes progress observable without drowning anyone in dashboards. A handful of signals tells you whether your measurement and optimisation culture is working; a simple maturity view shows where you are today and what “better” looks like next. Keep it light. Review regularly and use the trends to steer decisions.
The signals that matter
Time to insight
How long it takes from a user action, campaign launch or experiment ending to a meaningful insight being shared. Slow turnaround means analysts are overwhelmed, data is fragmented or tools are misconfigured. Fast feedback loops fuel better decisions.
Data quality confidence
The proportion of events, attributes and records that pass validation rules. Missing or inconsistent data erodes trust. A high confidence score indicates that consent, schema and governance are being respected.
Experiment velocity
The number of controlled tests (A/B, multi‑variant or personalisation trials) run per cycle and the proportion that lead to action. Too few tests suggest a risk‑averse culture or a lack of tooling; too many inconclusive tests indicate weak hypotheses or poor measurement design.
KPI improvement
The trend in your agreed KPIs—engagement, conversion, retention and revenue—over time, especially in relation to experiments and optimisation efforts. A healthy signal shows incremental gains tied to specific changes rather than random spikes.
Adoption of insights
How often decisions—product, marketing or content—explicitly reference data or test outcomes. Low adoption means reports are produced but ignored; high adoption shows that evidence is guiding roadmaps and budgets.
You do not need perfect tooling to start. Use timestamps in analytics for time to insight; validation logs for data quality confidence; experiment run logs for velocity; KPI dashboards for improvement; and meeting notes or roadmaps for insight adoption. Refine your measurement as you mature.
How to read the signals together
Fast insight, high confidence, improving KPIs – Measurement is supporting decisions; invest in more sophisticated experiments and predictive analytics.
Slow insight, poor data quality, flat KPIs – Fix your instrumentation and governance before scaling tests; the problem is upstream, not in the marketing mix.
High experiment velocity, static KPIs – Hypotheses may be weak or tests underpowered; review your experiment design and prioritisation.
Low insight adoption despite good data – Culture is blocking improvement; focus on storytelling, training and aligning incentives.
A simple maturity view
This framework isn’t a certification; it is a shared language to describe where you are and what to aim for next quarter.
Ad hoc
Data is collected sporadically; metrics are unclear; insights arrive too late to matter; experiments are rare and unstructured; decisions rely on anecdote.
What changes next: agree a handful of KPIs, standardise basic instrumentation and run a small test with clear success criteria.
Defined
Metrics and KPIs are documented; data flows into a single dashboard; experiments happen occasionally but follow a repeatable format; some decisions cite data.
What changes next: improve data quality with validation and consent checks; increase test cadence; schedule regular insight reviews.
Managed
Instrumentation is consistent; data quality is monitored; experiments are part of the delivery cadence; KPIs show steady improvement; decisions regularly reference reports; teams share learnings.
What changes next: shorten time to insight by automating analysis; scale personalisation experiments; link measurements to commercial outcomes.
Optimised
Measurement is continuous and predictive; AI assists with segmentation and forecasting; experiments run autonomously with dynamic allocation; data drives real‑time personalisation; insights are adopted across the organisation.
What changes next: sustain the rhythm, refine predictive models, and share the practices with adjacent teams and partners.
Keeping it lightweight
Put these five signals and your current maturity rung on a single page with a sentence or two on what improved and what you will try next. That is enough for leaders to steer and for teams to act. The aim is not more reporting; it is clearer choices about where to focus effort in the next cycle.
Workshop template
Get access to the Miro template to use with your whole team to work through the DXPlaybook
Glossary
KPI – A key performance indicator is a quantifiable measure that tracks progress towards a business outcome (for example, lead acceptance, conversion or retention).
Leading and lagging indicators – A leading indicator is a forward‑looking metric that signals future performance (such as engagement or intent), while a lagging indicator measures realised outcomes (such as revenue or completions).
Experiment – A controlled test designed to evaluate the impact of a change by comparing outcomes against a baseline or control group.
A/B test – A type of experiment that shows two versions of a page or feature to different audiences to see which performs better on a defined metric.
Data layer – A structured object that standardises event and attribute data, making it available for analytics, marketing and personalisation systems.
Conversion rate – The percentage of visitors who complete a desired action, such as submitting a form, registering, purchasing or downloading.
Attribution – The process of assigning credit to marketing channels or touchpoints for their contribution to a conversion or sale.
Engagement – A measure of how actively users interact with a site or content, typically including metrics such as time on page, pages per visit and interactions with calls to action.
Time to insight – The elapsed time between an event (for example, a release, campaign or experiment ending) and a meaningful insight being delivered to decision‑makers.
Data quality confidence – An assessment of how complete, accurate and compliant your data is, often expressed as the percentage of events or records that pass validation and governance checks.
