Marketing’s not a technical field, right?
Think again. As growth hacking has proven repeatedly over the last few years, a data-driven approach to marketing achieves stronger, more reliable results than traditional marketing.
The mad men of the future will combine creativity with rigorous analysis. (tweet this)
Despite a plethora of outstanding tools that handle much of the analysis, the best marketers will understand enough statistics to explore deeper insights from Google Analytics, enough SQL to access product usage logs without waiting for IT to build a special interface, and enough programming to effectively communicate requests to engineering.
Learning these seven technical skills will help you grow your product faster and more effectively.
Much of the effectiveness of growth hacking comes from its data-driven approach to marketing. As such, a basic knowledge of statistics will help any growth marketer become more effective.
Now, this doesn’t mean you need to go back to school and major in applied stats, but it does mean that investing a few hours a week learning basic statistics could be worth the time.
To run a/b tests well, you should at least understand the basics of confidence intervals and significance. To forecast growth patterns, you should at least understand regression- and smoothing-based models (plus data visualization). To mine insights with SQL, you should at least understand basic statistics and logical structuring.
Resources For Learning:
Challenge Project: Predict Crime
Use publicly available data for your city to create a predictive algorithm (check for your city here). Based on all the historical records, build an algorithm to predict crime by type and location (probably zip code). To check for accuracy, you could build the algorithm off of data up to 12 months ago, then test it’s accuracy in predicting this past year’s activity.
For an easier project, you could use that same data to create a safety rating for each part of your city – perhaps visualizing it in a map. (like this map modeling energy use in NYC)
While you shouldn’t necessarily become a programmer, all growth marketers should learn the basics of coding. It will help you build small things yourself (like a/b tests) and let you communicate more effectively with your company’s engineering teams.
In general, you should master HTML/CSS and have some familiarity with these web languages:
- Ruby on Rails
Remember, you don’t need to become an expert programmer who can rebuild Pandora from scratch – you just need to know enough to read code that’s already written and possibly make a few small modifications.
For instance, I’m certainly far from a PHP master or even novice, but I do know enough that I can generally find the right code to add or delete within my WordPress theme files.
Resources For Learning:
Challenge Project: Build A Personal Website
Particularly for people working in digital marketing, personal branding though a portfolio website will supercharge your career. You don’t need some super complicated website or even a blog, but you should have some central place people can visit to learn more about you (example and example).
Using WordPress as your website framework, use coding (HTML, CSS, and PHP) to customize your website’s look and feel. Depending how ambitious you’re feeling, you can start with either a free or paid theme, then learn as you go until you like the way your website looks. Wordpress is a great framework to use for this because the WordPress support community is massive. When you hit snags in your coding, the solution is probably just a search away.
3. Excel Modeling
Big data. Everyone wants big data, but no one knows what to do with it. The elephant in the room, of course, is that big data is absolutely useless by itself.
To be useful, big data must lead to big insights. (tweet this)
And that’s where Excel modeling comes in. Once you have data about how your customers use your service, you should be able to throw that into a pivot table and uncover interesting insights. Perhaps customers who downloaded ebook 2 are twice as likely to join as those who downloaded ebook 1. Perhaps people who got both are even more likely. You’ll never know unless you can turn the raw data into key learnings.
Happily, you can learn enough Excel to be dangerous pretty quickly by focusing on three key elements: formulas, pivot tables, and macros. Keyboard shortcuts will also come in handy.
Resources For Learning:
- PivotTables 101
- Introduction to Excel Financial Modeling
- Udemy’s Financial Modeling and Valuation Course
Challenge Project: Mutant Cell Growth
Your neighbor, a top-secret experimental scientist, has noticed a disturbing number of mutant cells in his lab. Knowing you’re a data ninja, he asked you to forecast mutant cell growth.
Your scientist friend currently has a total of 100,000 cells in his lab and 1% are mutant. Every day, each cell replicates – producing 15 normal cells and 1 mutant cell. Survival rate is low, though, and only 1 out of each 16 new cells randomly survives.
At the end of 10 days, how many (1) normal cells, (2) clones of original mutants, and (3) new mutants will your neighbor’s lab contain?
Sure, you know your total visitors and you’re seen the colored map of visitor locations, but do you know whether people are more likely to buy from an Android running Chrome or an iPhone running Safari?
The real power of analytics comes in segmenting visitors (by platform, country, source, demographics, etc) and discovering the outliers. Some of these insights will result in major conversion rate boosts when applied to everyone while some will trigger boosts through advanced personalization.
Resources for Learning:
- Official Google Training
- Distilled’s guide to mobile tracking
- Distilled’s guide to custom variables
- Mixpanel University
Challenge Project: Track Phone Conversions
You’re not just a normal ecommerce company that can use Google Analytics’ default conversion tracking. Instead, you’re a local pest control chain, and most of your customers convert over the phone.
How will you track conversions from your advertisements online through to your conversion on the phone?
You’ll need a call tracking service (my favorite’s CallRail). Most of those services make it easy to record phone calls as conversions in Google.
But what if you want to record your real conversions – when customers actually pay? That’s a little more difficult, but still possible: You’ll need to write a custom script that exports from CallRail’s API and imports through Google Analytics’ API based on user cookies.
5. Database Querying / SQL
GUIs like Google Analytics and KISSmetrics make it easy to do basic analysis, but sometimes you need to really get into the weeds and study your user behavior.
If you ran Facebook, for instance, you’d want to know how often users edited their profile, how many new pictures they updated, and how much time they spent browsing the newsfeed. That information wouldn’t be in Google Analytics, but it’s probably tracked in a database somewhere.
Using SQL to query that database will give you powerful information to move the dial. Perhaps newsfeed reading time is directly correlated to recency of content.
You’ll only uncover these insights if you know how to query raw databases with SQL.
Resources for Learning:
Using the sample databases provided by W3Schools, find the top three employees based on revenue generated.
To do this, you’ll need to join a few of the tables provided to determine the value per order and from there the revenue per employee.
You could solve this quickly by simply exporting data to Excel, but force yourself to do it through one SQL query. SQL can handle much larger data sets than Excel, so you may as well practice using SQL from the start.
Hint: Most questions you encounter can be answered here or with a quick Google search.
6. A/B Testing
We’ve all seen the power of a/b testing, but how many people really understand how it works?
Beyond creating a one-off button experiment in Optimizely, a good marketer should create a series of experiments designed to increase conversions and contribute to a better understanding of the target customer. Over time, only about 1/3 of all a/b tests run will likely contribute to an increased conversion rate – the other 2/3 should contribute to a better understanding of your customer.
Like most “magical” things, a/b testing isn’t really that complicated once you understand it. You simply need a document recording all previous tests (hypotheses, results, and screenshots of variations), a document recording all planned experiments, and a testing platform like Optimizely or Visual Web Optimizer.
Resources for Learning:
- Unbounce’s Ultimate Guide to A/B Testing
- VWO’s many case studies
- Avinash Kaushik’s Primer on Experimentation and Testing
Challenge Project: Create Your Own Testing Plan
Whether for your company or just for your own personal website, create a testing plan. Looking at case studies for ideas, create a list of all the tests you’d like to run and prioritize by potential impact. Then, as you test, carefully document the results of each experiment (and make sure you achieve statistical significance).
After running a few of these tests, review what you’ve learned about your target audience.
7. Web Scraping
More information becomes available online every second, but much is relatively inaccessible. Imagine the potential leads you could find and the optimization gems you could uncover if only you could scrape the internet!
Happily, you can scrape the web, and you don’t even need to learn programming!
With tools like Import.io, you can visually identify the information you want to harvest, and the scraper tool will find and aggregate similar data. If you’re working in a primarily sales-focused company, this is an invaluable tool. If your company is B2C, you can still use web scraping for everything from finding potential partners to analyzing your own blog.
Web scraping is one of those relatively simple skills that will put you light years ahead of the competition.
Resources for Learning:
- Web Scraping for Sales and Growth Hackers (my favorite)
- The Content Marketer’s Guide To Data Scraping
- Guide To Grow Your Blog 10x
Challenge Project: Find 10 Potential Blog Partners From GrowthHackers.com
You run growth at a b2b startup that sells tools for marketers. You know partnerships with bloggers can help drive acquisition, but how do you find the best partners?
Using web scraping, pull all the post, user, vote, and comment data from GrowthHackers.com. Then, analyze the data to find the most popular blogs that could be potential partners. If you’re looking for individual influencers, you could do the same thing, but just analyze the users instead of the blogs.
For bonus points, you can feed all the URLs you scrape through SharedCount to get the social sharing stats also.
Your results will probably look something like this.
Most industries have their own niche forums and websites, and you can leverage scraping to uncover the influencers in your market.
The vast majority of people reading this post will do nothing. They’ll nod wisely, perhaps share the link on Twitter to impress their friends, and continue to the next post in their feed reader.
A few people will try to do everything. Eagerly jumping in head first, they’ll try to master all these technical skills simultaneously, but within a day or two they’ll give up – overwhelmed by the sheer volume of work they’ve heaped upon themselves.
An even smaller group will pick one skill to start learning today. Following the principle of serial excellence, they’ll focus on that one goal until they finish, then move on to the next.
Reading articles about growth is a good first step, but the real rewards come when you start taking action.