Superweek Hungary Analytics

Twitter: @iProspectUK

Superweek.hu, GalyatetőJanuary 23rd, 2014

ON THE ORIGIN OF DATA, BY MEANS OF IMPLEMENTATION

OR INSIGHT

Rachel Sweeney


The Origin of Data


Driving Digital Performance


THE INDUSTRY THAT WE ARE IN HAS BEEN EVOLVING




2011-12: WEB ANALYSTS STARTED TO BECOME ‘MAINSTREAM’




2013: CONTAINER TAGS AND ANALYTICS UPDATES CREATE FLEXIBILITY




2014: AN ACTION PACKED YEAR OF BIG DATA INSIGHT




MORE DATA DOES NOT NECESSARY EQUAL MORE TIME FOR ANALYSIS

• Only 24% of clients could confidently say that their budget for analysis have increased

• Analytics departments are growing… as are the number of clients/sites managed

• Compromise is not a buzzword.




HOW DO WE GET TO A DATA DRIVEN DIGITAL MARKETING NIRVANA?

Our goals:

• Spend more time on analysis less time on implementation and reporting

• Provide insight to all parts of the business

• Increase business profitability not just website performance

Always on Campaign

Data

Strategy

Implementation

Increase Business Conversions




ACTIONABLE INSIGHT… WHAT DOES IT LOOK LIKE?




HOW ARE WE GOING TO START DELIVERING ACTIONABLE INSIGHT?

• The reality is that majority of companies are not in a position to start delivering actionable insight tomorrow.

• It is a process of drip feeding data trying, testing, seeing, learning, and adjusting

• Invest in a statistician/mathematician.




ACTIONABLE INSIGHT DEMANDS INTELLIGENT IMPLEMENTATION

• To understand evolution, you must understand the origin

• Now is not the time to start from the beginning

• Now is the time to assess that the setup that you have, is most likely to support the future.




HOW DATA INTELLIGENCE HAS EVOLVED AT IPROSPECT

• Department has grown from 2 to 11 in 2 years• Hired a statistician and server-side developer

• Integrated operating companies to provide clients with one agency, delivering specialists in all areas.

• Analysts is less ‘web(site) analytics’ and more, the data unit within strategy, implementation, and insight teams.


The 80 | 20 model




HOW TO ACHIEVE EFFICIENT INSIGHT FROM THE ORIGIN OF DATA

• Generic tracking

• Specialist

• Bespoke• Evolution

Business Objectives Integration

AutomationPlateau

80 2080

80802020

20

Twitter: @iProspectUKTwitter: @iProspectUK

“80% of tracking requirements should originate from the core

business objectives”




TWO SCHOOLS OF THOUGHT: TO START AT THE BEGINNING OR THE END?

Let’s tag everything

Let’s define KPIs

Beginning End




AT IPROSPECT WE ARE IN THE ‘PRO-END TO BEGINNING’ CAMP

TEAM KPI*

*However we always track a few other cool things




THE 80% INCLUDES WEBSITE AND CAMPAIGN DATA REQUIREMENTS

• Bucket your website’s data requirements: – Site usability and functionality– Site Performance– Campaign Performance

• List your KPIs and audit what data is needed and from what source. Also list your stakeholders, client, agency, internal, external, names, positions, experience.




WHY SHOULD WE FOCUS 80% ON BUSINESS REQUIREMENTS?

• We need to make our implementation efficient to be able to focus on insight

• Focussing on obtaining data that will directly provide insight to be fed to the strategy team is the most efficient way of getting to the ‘actionable insight’ stage

• This doesn’t mean we limit ourselves

• Growing the business is the ultimate goal





• Generic tracking

• Specialist



AutomationPlateau

80 2080

80802020

20


“Actionable insight is not limited to website data so neither should

your implementation be”




80% INTEGRATION, 20% SPECIALISM

• A website’s setup is one part in a landscape of data.

• It is more and more relevant to ensure that all data sources are aligned and complement each other

• This will provide consistency when analysing




INTEGRATION IS NOT SOMETHING YOU CAN DO ON YOUR OWN

• Visit the teams • Work on a process that suits all parties• Create boundaries• Continue relationships and networks




CONTROLLED CONSISTENCY IS THE MOST IMPORTANT QUALITY SETUP

• So many things that cannot be consistent, it is so important that changes are made to become consistent.

• Naming conventions, account structures, campaign separations must all be as consistent as possible

• Regular (automated) auditing will show where inconsistencies are as inconsistent data can devalue insight.






• Specialist• Generic tracking


AutomationPlateau

80 2080

80802020

20


“An analyst must do 80% of the work in 20% of the time, then another 80%

in 20%, then another…”




ACTIONABLE INSIGHT RELIES ON AUTOMATED IMPLEMENTATION

• Businesses want to see data and conversion rate increase but aren’t necessarily increasing time on insight

• As analysts we need to be as efficient as possible in implementation

• Automating implementation is absolutely key to getting to the ‘actionable insight’ stage




TAG MANAGEMENT AND APIS FACILITATES AUTOMATION

• Get to know the management API

• Understand tag management systems, their approaches, their limitations

• Design your tags to be as efficient and consistent as possible. Intelligent tag management implementation drives intelligent data




SOME INSIGHTS CAN BE AUTOMATED

• Get to know the reporting API

• Create dashboards that don’t just report

• Define what regular insight may be useful and create mathematical models/ quick and dirty spread sheets to free time

• Insights from good data are far easier to pull than insights from problematic data






• Specialist• Generic tracking


AutomationPlateau

80 2080

80802020

20


“Insight is not ‘perfect’. Neither does your implementation have

to be.”




DON’T WAIT FOR MRS/MR RIGHT, WE NEED MRS/MR RIGHT NOW.

• Providing small insights lets the key stakeholders know what information is available and lets you know what is useful to yourself

• Don’t presume that the receiver knows what insight they need.

• Follow an A/B testing approach to providing insight. Create surveys internally. Apply the same optimisation you perform for websites, to your teams.




PUT DOWN THE TAG MANAGER, AND PICK UP THE DATA VISUALISATION TOOL

• Invest in other data visualisation tools

• Hire individuals who have a background in stats

• Now is the time to re-do the way data is visualised, not the way you track

• There is a variety of good visualization tools including Tableau, Google Charts, Google Visualization API, Highcharts JS


Intelligent implementation will result in 80% of time spent on

insight.


Questions?

Rachel SweeneyiProspect [email protected]

Superweek Hungary Analytics

Technology