What are Analytics?
- analysis of data
- i.e. analysing your business and competitions's data to drive continual improvement
Recent trends in Digital Analytics
- Big data
- Cloud
- Diagnostic analytics
Data
- numbers, words, opinions
- for digital, mainly numbers
- Qualitive data is gathered through user research (observing people to understand why they do certain things)
- Quantitative date through analytics (identifying what actions users take when they come to a page, how many users take those actions)
Quantitive data allows us to measure baselines which we use to inform design decisions.
Analytics is for everyone and is everywhere.
An analyst cannot simply measure hits on server, must measure human behavious.
- likes on facebook
- likes on tweet
- steps in a day
- 2.4 billion people connect and interact in one minute
- 80% of consumers are more likely to recommend a brand that offers a simpler experience
- 80% - 90% of companies consider big data analytics top priority
Cloud computing makes it easy to store/collect data.
FOUR DIFFERENT TYPES OF ANALYTICS:
- DESCRIPTIVE ANALYTICS
- What has happened?
i.e. How many customers bought a chocolate bar last week? - DIAGNOSTIC ANALYTICS
- Why did it happen?
i.e. Why did that many people buy a chocolate bar on this day? - PREDICTIVE ANALYTICS
- What will happen?
i.e. What will happen if I run a buy one get one free campaign on chocolate bars? - PRESCRIPTIVE ANALYTICS
- How can I make something happen?
- How can I double the sales of my chocolate bars?
These types of analytics use metrics based around Key Performance Indicators (KPIs).
KPI - measureable action/signal that is correlated to business success.
Retweets on Twitter don't directly increase how much users like or know and organization. Ideally, an organization should have multiple KPIs for one business object to increase reliability of data.
COMMON METHODOLOGIES
Research
- prioritize goals/questions in order to focus attention
- once know goal, create hypothesis then test it
- data analysts measure results of research and tests
- researchers and analysts recognize outliers
- come to conclusions, able to predict future outcomes based on patterns they identify
Measurement
- most metrics help us understand how an organization/brand is growing
- measure number of users, speed of sites. time spent on page
- offline details such as amount of money made, number of sign-ups, number of purchases
If you only measure without research, nothing to compare to, data is meaningless. Data tracking is ongoing measurement supported with reasearch with an intent of analysis.
Analysis
- break down infromation into smaller pieces, examine what it means
- gives meaning to measured data
- allows us to make connections
Daily Tasks and Deliverables
- setting key performance indicators
- optimizing content
- setting up analytics tools
- monitor and measure
Tools of the trade
- Google analytics
- Moz Pro
- Clicktale
- KISSmetrics
- Crazy Egg
How Analytics are used to supercharge your business
- Targeted direct marketing
- integrate customer data from multiple web/social media interactions
- determine effectiveness by customer type, location, delivery channel - Predictive advertisement targeting
- present ads that consumers want to click - Fraud detection
- Need to knwo which transactions/applications for credit, benefits, reimbursements, refunds etc are fraudulent.
- businesses need to minimize false insurance claims, inaccurate credit applications, false identities - Investment risk management
- 'big data' needs evaluating with predictive analytics in order to know what to invest in - Customer retention with churn modeling
- every business wants to know if customer is about to leave and why, so they can do something to sway them to stay
- learn what incentive one-time customers need to return - Movie recommendations
- movies recommended based on past reviews, related interests, analysis of twitter comments etc
- analysis on movie scripts, based on reaction to films, predict box office revenue - Education - guided studying for targeted learning
- guidance on which question areas need more study
- every student needs help on how to spend their study time more effectively
- schools need same analysis to provide effective teaching - Political campaigning with voter persuasion modelling
- find out what persuades voters phone call, door knock, flier, tv ad) - Clinical decision support systems
- costs escalating, important to determine which patients are at risk of developing certain conditions
- predictive analytics can help make best medical decisions - Insurance and mortgage underwriting
- allow auto insurance companies to accurately determine cover aswell as their bottom line
- assess borrower's ability to pay back mortgage
IMPORTANCE OF MEASURABLE OUTCOMES
- identify business problem, then understand how it can be measured
Five objectives useful to measure for business:
- Branding
- Online information and support
- Content publishing
- Lead generation
- E-commerce
Measuring metrics for business success
The measurement Process
- Create a strategy
- at least a general idea of what you want to accomplish - Define Key Peformance Indicators
- ensures discussion around metric focuses on how to improve whether than debate if it's the right metric - Implement them
- implement KPIs prior to launch to get the most data to use - Measure and Improve
- Baseline from launch
- Baseline against historical average
- Baseline against the same period last year
Conversions
- - key metrics in website analytics for online businesses/blogs
- an activity on your site that is important to the success of your business
- better to know how visitors engage with your site (click link, purchase product)
Measure conversions
- conversion rate (no. of conversions divided by number of visitors)
- higher the conversion rate, higher the engagement with end users
- to increase a low conversion rate, maybe look into changing website layout/placement of products
Types of conversion
- Micro conversions
- activities that precede a purchase on a website/business objective</br> i.e. signups for an email newsletter/account creations - Macro conversions
- general idea of how website is performing
i.e. how many sales transactions for a specific product, number of online insurance quotes
Macro conversion is number of users following a blog. The macro conversion is purchasing something you talked about on the blog.
The continual improvement process
- consists on gathering data, reporting, analysing, testing, insights
Example with chocolate bars
- understand how many choclate bars were sold
- we report on it
- analyse the report ( why sell that many? What were drivers?)
- Testing (offer on bar? change packaging?)
- Analyse results of the testing to get insight (price is key?)
Analytics is an iteractive process.