Growth Hacker and Data Punk Daniel Cave talks about how how to put Data at the heart of your business.
What you should track, what you should share and who you should share it with to drive the best business decisions possible.
1. Analytics and Data Culture
Lessons from a fast growing tech SaaS startup
Putting data at the heart of your business
to make fast decision based on data
2. Quick ‘about me’
SEO / Technical Import.io
• Technical Marketing/Growth Guy
• In Product Analysis and Stats to drive insight
• Contractor Intuit, Wordtracker, linkedex User.
6. “If you are not collecting data now, its
impossible to collect it later!”
So collect all data even if you are not using it
right now. Try to plan ahead, and imagine what
you will need to analyse in the future.
Lesson #1
7. With Any Dataset, Perspective and Context are
paramount. Who has the domain knowledge for
each data point?
Lesson #2
8. “Not Putting Data culture into the heart of your
business and not including everyone makes for
worse decisions”
Lesson #3
9. My Philosophy
Gather / Store / Display - Bits of Lean
• Learn to Collect Everything
• Learn how to Share the Right things, Right
People
• Make better choices based on data viewed through
your domain experts (your team).
12. Culture: What we Track
Everything, the whole journey for every user.
• Acquisition
• Signups, Browser Download, Video Played, Contacts,
Forms Submitted, Content Type Consumed, Form Fields
Clicked, Campaigns sending traffic.
• Activation
• Successful and failed uses of every feature, support
requests, enterprise data requests.
• Retention
• Returning Visits, Regularity of visits
• Active and Core Users
• Which customers are REALLY engaged
22. “Everyone wants to have the voice and
opinion heard…
Everyone of your team has a unique view
on what the data could mean…
Everyone can see what actions could
influence it.”
25. How we use/share the Data
Measure Real Success!
• Develop & Analyse Core User Profile Using Data
• where a user came from (acquisition trail)
• when they visited
• what device they visited on
• how/when/why did they signup
• which features they used
• what task they tried to complete
• how many times it took them to complete the task
• how many did not complete
• how many times have they visited in the last week/month
* Cost per Goal Completion
UI / UX + Analytics
Why?
Marketing+Web Analytics
Why?
User Success + Analytics
Why?
26. How to use that Data
Segmenting and analysing as a team.
• Marketing: Knowing your Core User CPA by channel gives
you solid ground to advertise from.
• Product managers and Developers: Test new features
for effect on retention, adoption and profit.
• Social/Engagement: Engaging People who are already
bought into your business (aka core users) reaps big
dividends.
• Support: Squash common issues as a priority and support
your theory with data.
27. Examples
Simple Cases
• Segment and Email Core Users Using SQL + Woopra,
then invite them to your ambassador program.
• You Can Get
• Free content from them
• Free Social Media exposure
• Free Links from there Network
• They Get
• Freebies
• Insider emails
• Advanced webinar tickets
28. Examples
Simple Cases
• When you implement a New Feature
• Have visibility on its real world effects
• Does it increase the percentage of Signed up to
‘activated’ users (aka real growth).
• Does it effect the CPA of core users and/or
retention rates.
• Have you reduced or moved the “time to Wow”
29. Examples
Simple Cases
• When you are buying visitors/media
• Have visibility on real deep ROI
• Is it sending visitors that stick?
• Exactly which visitors did it send and what did
they do different to others.
• What is the CUCPA*
• Improve your persona to product fit, which traffic
sources drove use of (or purchase of) which
features.
31. The advanced stuff
…for another day
• If you have the right data you can start creating a
predictive analytics model to help plan spending and
buying.
http://en.wikipedia.org/wiki/Regression_analysis
What is unique about my situation that lets me give you a different angle on Analytics?
The answer was pretty obvious… import.io is a super high growth Saas product in data space.
Why is that a good thing?
There are many parallels that can be drawn between growing a Saas product and any website, we just have a slightly different perspective. where as a want to acquire users you may want to acquire readers, customers, or subscriptions.
The other difference is, I work with data everyday;
We are providing datasets to enterprise size large and small and the one thing I see them ALL doing is innovating to make faster smarter decisions based on data intelligence… So my thought process was “how can I Help you do that too with improved, but easily implementable analytics practices”… and so i came up with this presentation.”
I sit between Marketing and Product
The sorts of things I get up to:
SEO - covering a multitude of jobs from SEM, to A/B testing for CRO, Outreach, and display media buying etc…
Im also one of the ‘Data Guys’ or ‘Data Punks’ as we call them. We are the champions of using data in decision making across marketing and product.
What I want to share with you is…
There are three things to talk about today.
…And the thing which makes any of this any use at all is… Data Culture
There are three things to talk about today.
Leading metrics:
These are what is important to ultimately change in a positive direction day to day. They are the easy to influence metrics that have large effect on day to day working.
If you want to see a spike in these numbers, get into life hacker, or an industry mag.
These should be shown constantly. They WILL have a psychological effect on individuals, focusing and showing progress (or lack of it) transparently removing procrastination and encouraging action. They help focus efforts on where they need to be to meet short term objectives and key results.
Lagging metrics:
These are the metrics which are more typically stubborn but often the roll up result of the leading metrics. A good example is growth rate… if you increase Signups and Conversion rate, via better UI, you will see as a result week on week the growth rate increase or CLV/CUCPA drop.
Leading Data can give immediate feedback band keep teams focused on the tasks at hand.
lagging data will give you clarity on progress and business performance.
They can all be useful but while some people should focus on leading, someone HAS to be focus on the lag.
There are three things to talk about today.
Data Lunch is a time when a leader bring there data analysis to the table and invites discussion.
Context and domain knowledge transfer is the name of the game. Invite support, sales, marketing, leadership and all other departments in to see the data, and to provide you with the deep insight you can use.
We dont stop there:
once data has been shared we all meet for an hour on a friday afternoon… beer optional… and hash out our theories and ideas around what could be done to improve the product.
These ideas are all measurable and are all based on data observations.
Your Users are your ultimate domain experts. If you want to make something better for a user, ask them how to do it. Often they provide us with qualitative data to investigate quantitatively.
We ask them for feedback and we get it!
We let users vote on ideas and feedback too, so we can collect data on this… it even lead to a pivot in a fairly large decision around Auth APIs.
Deep segmentation can happen if you know,
where a user came from (aqusisiton trail),
when they visited,
what device they visited on,
how/when/why did they signup,
which features they used,
what task they tried to complete,
how many times it took them to complete the task,
how many did not complete,
How many times have they visited in the last week/month
Core User CPA needs to be calculated from your metrics:
Our main Core user Definition revolves around “use recency and API access count".
Your main Core user Definition may revolve around how many shares a user does for you, and how votes they give to a story.
support is an interesting topic for Analytics… tagging and categorising your support tickets and support page views can help you understand which issues are effecting users the most… a vital thing to know if you are going to help them succeed.
Cost per core user acquisition
Which traffic sources drove use of (or purchase of) which features. EG: did Data journalists use the crawler more, compared to Developers who used a high percentage of Connectors?
Do Growth Hackers access the data they structure via API or the Downlaod button… and how many try?
Another great thing about having this data is that you can run more advanced analysis on usage and acquisition. Her is how we monitor a vital factor for any site, how often a user comes back for another visit.
A co-hort analysis is important to see how groups of people within a time frame were effected by developments in your site.
If you image this graph shows the number of people who were active on your site having ‘signed up’ on day 0. If its important to keep your users engaged for just one week or ten, you need to pay attention to this kind of data.
Lets say you sell online courses which people complete onsite behind a login… your primary goal should be to keep as many people coming back as long as possible so they can complete the course and buy more stuff…
You can also look into each slice of this data to reveal the types of activity which are happening between each slice.
in this example there are two big drops in returning people… because we track everything down to the individual user journey level we can now see what people were trying to do at these stages, which features they were using, which features did they try to use and fail on… all good information.
In this instance, in Jan, we seen 49% of people who signed up didn’t even try to use the tool. Of the remaining 51% who did try to tool in week 0, 27% (61.5%) came back, which is a 38.5% drop off!!!
You can run these calculations for every step, figure out insight into the problems your users are having, and figure out how to do a better job. then test those theories.