2014 is almost here. How will you make more money in the coming year? Here are my 10 rules for winning in 2014.
This post is based on my keynote presentation at Mesh Marketing last week.
Question: what do these 4 things/people have in common?
- The awesome book (and mediocre movie) Moneyball
- Barack Obama
- Jeremy Lin
- Nate Silver
- “Moneyball” was a story about a mediocre, under-funded baseball team achieving great things when managed by looking at the data.
- Barack Obama won the 2012 US elections and his data-driven team played a huge part in it. Among other things they conducted 240 A/B tests that resulted in a 49% increase in donation conversion (that was more than $80 million in extra cash).
- Jeremy Lin didn’t appear out of nowhere and become a good player for the NY Knicks. A statistician predicted the success of Jeremy Lin 2 years before anyone knew his name.
- Nate Silver became famous for predicting Obama’s wins in the last 2 elections based on his analysis of polling data.
These are four cases when a data-driven approach kicked traditional thinking in the butt—hard. And you know what? These events occured in 2012, excluding “Moneyball” which came out in 2004. This happened in the past and is nothing compared to 2014 and beyond. If you’re not data-driven, you’re stupid.
This leads me to rule #1.
#1. Don’t be stupid
It pays to be smart, it really does. That’s obvious right, so no problems with this? After all, 91% of CMOs believe that successful brands make data-driven decisions, per Columbia Business School. However, only 11% of marketers use data to make business decisions today. Maybe data-driven stuff is not all that useful? Well, a Nucleus Research study that showed that an incremental 241% ROI can be generated by applying data to business decisions.
We have proof that data-informed decisions provide a greater ROI, yet most companies are not using data to make decisions. Smart or stupid?
So what are the other 89% doing? Typically instead of data, they are driven by HiPPOs – highest paid person’s opinions (credit for the term goes to Avinash).
Now let’s look at another piece of data. Research shows that marketers are spending pennies on their conversion efforts. For every $92 spent to acquire traffic, just $1 is spent to optimize it. (The data is now a couple of years old, so the gap has probably shrunk, but still!)
So yes, companies spend ALL that money to get traffic to their site via SEO, PPC, and social media and then throw the traffic into the garbage, wasting all that effort and money only because their website is not optimized for conversions.
Conversion optimization is pretty much the only digital marketing activity that is 100% focused on making more money. And you only spend $1 on it compared to $92 for traffic? Please stop being stupid in 2014.
Not happy with your conversions?
Recent Econsultancy conversion report said that only 28% of marketers are happy with their conversions. Well, no shit! Remember 92:1? Yeah, that has probably something to do with it! My guess is that the 28% that are happy probably don’t know any better.
Fortunately, 59% of businesses now say conversion rate optimization is crucial to their digital strategy. Most of your competitors are starting to figure this stuff out. If you’re still living in the 92:1 world with your budget, you’re gonna be left behind.
The thing is that SEO is going to get increasingly harder. If you’re still doing SEO by optimizing for keywords, you should know: it’s not 2008 anymore. Things have changed! Google understands context, natural language, intent.
What about PPC? If you’re investing in PPC you already know the costs have gone through the roof! It’s very hard to buy traffic profitably. So no wonder 59% of businesses have figured out that they should fix their conversion machine before pumping any more dollars into their PPC or SEO campaigns.
Still not convinced?
Let’s play a little game. Lets say you run an ecommerce store with an average transaction value of $50. And you get 100,000 visits per month to your site. So this would be the difference in your monthly income based on your conversion rate:
1% – $50,000 / mo
2% – $100,000 / mo
4% – $200,000 / mo
8% – $400,000 / mo
12% – $600,000 / mo
The difference can be huge. Think hard about that $1 investment again.
What about just implementing all the best practices out there? Ha! Remember the old marketing adage “half of the advertising dollars are wasted, just don’t know which half”? Best practices are similar. They only work on half the websites, and using a best practice is like using somebody else’s solution to somebody else’s problem. So yes, read best practices to get inspiration but customize it to your reality.
What about copying (successful) competitors? Don’t. They don’t know what they’re doing either.
In conclusion, if you make decisions based on these inputs, you’re stupid:
Time to change that.
#2. Learn how to win over the girl/boy
Recognize this movie?
Yes, it’s Groundhog Day! If you haven’t seen it, go watch it now. It’s the greatest split testing movie of all times!
Bill Murray lives the same day, Groundhog Day, over and over again, and he has a goal – he wants to get the girl. So he tries hundreds of different ways to approach the girl. Until he succeeds:
So what happened? How did he finally manage? He wins her over by figuring out he needs to show emotions.
The same scenario happens on your website. You need to speak to your audience in a way that resonates with them. And you can’t know the right approach in advance – you have to figure it out!
Let’s look at this example. Not bad at first glance? Well…
- “Welcome to our website”. – Complete waste of space that doesn’t resonate with anyone. Everyone is welcome on every public website on the internet.
- “Who We Are”. – Who cares! People care about themselves and THEIR problems. Not you.
- “Our Philosophy”. – Seriously? No one gives a damn!
- Cheesy stock photos. – So fake I want to flee instantly.
Instead of this crap, always be user-centric. WIIFM (What’s in it for me) is a great principle to keep in mind at all times. Your first order of business is to create relevant and interesting content. Then you have to present it well. Otherwise, you won’t win over the girl.
How to get great hypotheses for copy/content
Qualitative research is your best friend—it helps you to get to know your audience and observe the voice of the customer.
- Send out customer surveys to people who’ve just bought from you to learn about their experience and thought process
- Use exit surveys on your website to ask visitors who are showing engagement, but are not clicking the ‘buy’ button, why they aren’t purchasing (e.g. “what’s holding you back from making a purchase today?”)
- Call your actual customers and learn how your product fits into their lives.
- Talk to your customer support people and inquire about the FAQs.
- Do user testing on every new/updated portion of your website. It’s so cheap these days with tools like usertesting.com.
This post comes with specifics on figuring out your online audience.
#3 Understand that it’s about the process, not tactics
Your website is a bucket – and it’s leaking money. Your job is to be a plumber – to find those leaks and plug them. You won’t succeed by trying random tactics.
Most companies think that a new, shiny bucket (website) will fix the problem. It might – depending on how bad your current site is (and whether you’ve hit a local maximum) – but it’s usually not a good idea. The thing is that when you radically change everything, inevitably some things will get better and some things will get worse. You won’t know what is good or bad unless you update a portion at a time. The bad things will cancel out the good things, so there may not be any change in the conversion rate. In fact, it may even get worse.
Chris over at WiderFunnel has written about it in more detail.
So what I’m saying is throwing out your leaking bucket and replacing it with a new one won’t probably solve your problems.
In order to fix those leaks, you need to know how to find them and plug them, and every case is different. This is why you need a process.
Most don’t have one
Econsultancy study says that only 26% of companies have a structured approach to conversion optimization.
The study also confirmed that companies that have a process grow their conversions 2x faster. I like 100% faster, don’t you?
So what are the rest doing? Throwing spaghetti on the wall to see if that plugged the leak. Reminds me of rule #1.
If you can’t describe what you are doing as a process, you don’t know what you’re doing.
– W. Edwards Deming
Are you familiar with the Lean Startup methodology?
The Lean Startup methodology is an iterative process for building new companies, products, and services. You start with an idea—a hypothesis about a problem. You build a minimal viable version of the product to tackle the problem and launch it to your market. You measure everything that happens to it, get useful data, and learn from it. Based on the learning, you create a new hypothesis, and iterate—over and over again.
The faster you can go through this cycle, the faster you will learn and succeed. It’s about validated learning.
Lean Startup is essentially conversion optimization—just a different name for it, but the idea is the same.
- Measure every single piece of data on your website —down to micro-interactions. You gather data from various sources: user tests, qualitative research, mouse tracking, etc.
- Take the data and analyze it, compare it with heuristic analysis findings, and identify problems to form hypotheses for a solution.
- Create optimized versions (treatments) for various pages using your findings and your know-how about design, UX, psychology, and so on.
- Test the new version against the control, and make sure you feed the test data to Google Analytics so you can segment the crap out of test results. Based on the results you get new insights and validated knowledge.
- Rinse, repeat.
#4. Know where to get the data
A while ago we were talking to a prospect. We were explaining the first step is conversion research, getting to know their audience, and figuring out what’s going on with their website. He wrote us this:
“I’d like to try and improve the actual conversion rates, instead of doing all the testing or gathering feedback.”
It’s like saying, “I don’t want to learn how to code, can I just get a top developer salary without it?” You have to do the heavy lifting to get heavyweight results. If you skip the research phase, you’re breaking rule #1.
There is no single ‘right way’ to do conversion research, but most successful programs include the following 6:
- Functional testing – Does your website work perfectly with every browser on every device? Lots of low hanging fruit here! How about site speed? Another easy one.
- Heuristics – No substitution for human-led assessment of websites. Of course, no random opinions, but a structured critique. Lift model is a great one to use.
- Web analytics – Google Analytics (or whatever you use) is a goldmine. Ignore vanity metrics, segment super deep and don’t just rely on built-in reports. Be ready to spend days digging in the analytics data.
- Mouse tracking – Where are people clicking, where are they looking, how far down are they scrolling? Secret sauce here is also watching user session replay videos, can be very insightful stuff.
- Qualitative research – User surveys, exit surveys, phone calls. A must.
- User testing & usability – User testing and making sure your site matches most usability standards.
So your design and other business decisions have to come from the right inputs:
#5. Become a master of analytics
The times when marketing people used to be the creative folk sipping coffee and talking about ideas is over. It’s never coming back. Yes, creativity is still important, but all of your design and campaign ideas are merely hypotheses that need to be tested. You need data to back your ideas up.
Data-driven design is far superior to somebody’s personal preference. In fact, make this a rule in your office: whenever somebody is about to voice an opinion, they have to start with “In my personal, unsupported, insignificant preference…”. This will set the right tone to a value of an opinion.
If you’re not going to be good at analytics, you’re going to be obsolete
Fast. Faster than you know. Digital marketing is going to be more and more analytics and data-driven.
Resources for learning:
- Digital Analytics Fundamentals course by Google. Free.
- Web Analytics 2.0 (book from 2009, but great foundation)
- Lean Analytics (book, mainly for startup folk, but good primer for all)
- Occam’s Razor blog
- Justin Cutroni’s blog
You have to know *everything* that is happening on every page on your site. Being a master of analytics will help you move from “I don’t know” to “I’ll find out” and that’s powerful.
Averages lie, go deep
Remember that averages lie. There are no average people. What’s the average of a 12-year old girl in California and a 50-year old man in Toronto? A sexually confused 30-something in Kansas! Then companies optimize for that target group. Breaking rule #1 again. Always compare data between segments and distributions.
Also, vanity metrics are not insightful. Whenever you’re looking at a metric, ask yourself: what am I going to do with this? If the answer is “nothing”, either figure it out or ignore it.
#6. Insight rules
Many people think that conversion optimization is about testing your button color, or some other such nonsense. Whoever says that is actually saying “I’m clueless!”. I get mad when somebody writes about a/b testing or conversion optimization and brings testing a button color as an example. On a side note, if you are wondering about which color converts the best, I have the answer right here.
So let’s imagine that you test these two buttons, and one of them converts better, you get a lift. Good boy.
So now what? What did you learn about your audience?
Best tests are based on hypotheses. You’re seeking validated learning. Once a hypotheses seems to work, you can build on it and optimize your site further using that same insight.
Here’s a landing page I’m working on.
We tested a bunch of headlines, each had a specific hypotheses behind it, and no matter what we tried we could not beat the simple, straightforward headline. This even beat the copy that was full of benefits—stuff the target audience was telling us they want in user surveys. So the insight we got from this is a simple, straightforward approach resonates with this audience.
- Cut down the amount of copy—and got a lift
- Made the page smaller—and got a lift
- Made the form simpler and shorter—and got a lift
… and so on.
What I’m trying to say here is that you should capitalize on your insights.
So what if a test fails to bring an uplift? A lot of them do. No worries. You segment the results and you’ll often find your losing variation was winning in one segment or another. Which one? That’s another insight for an improved hypothesis.
#7. Learn to avoid change and experiment instead
When I started in this business, I used to tell clients “based on our research and data, we need to change this and that on your home page” and got a lot of resistance. People don’t like change. Change is scary. I quickly learned to rephrase it as “let’s run an experiment”. Nobody is against experiments. This also works great to get buy-in from top management for testing.
You need to get everyone in the company on-board with experimentation.
Build a culture of experimentation
Get everyone to understand:
- There are no ‘holy cows’.
- Nobody knows what’s going to work.
- Cherished notions are stupid notions.
So everything should be up for grabs. Higher revenue and conversions are more important than the boss’ wife liking the sliders.
The more support you have from everyone to experimentation—IT, legal, finance, C-level people—the faster you can experiment. The one who runs the most experiments wins.
Get everyone involved
Make data about your website performance public within your company. Have everyone wanting to contribute to test hypotheses. Before making a new test live, have your staff bet—with real money—on which version is going to win and why. Once the results are in, analyze it together.
This gets everyone involved, interested, and excited.
Conversion optimizers should be free to run experiments fast, without needing C-level approval for each test, and they should get high priority from the IT department (those guys have to be involved, be part of the team, contribute). If marketing and IT is us vs. them, it’s gonna be trouble.
#8. Allocate resources to conversion optimization
88% of companies cite lack of resources—people and funds—as the key reason why they’re not doing more conversion optimization. Now that is clearly breaking rule #1.
Conversion optimization makes companies money. Yet they don’t allocate people and budget to it, and then complain about shitty PPC conversions?
Solving this is super easy. Take budget from PPC and display ads, and put it in conversion optimization. Done! After you get your site to convert better, you will make more money and can increase your ad spend again, bringing in more profitable traffic than ever before.
#9. Test it right
Test carefully. If you call a test early and run with false data, you’re making the whole process useless. Don’t rush it, make sure your test data is accurate!
- Watch out for any technical issues—sometimes reported data is wrong due to poor implementation.
- Test until 95% confidence level or above (meaning there’s only 5% chance that the result is a fluke).
- Make sure the margin of error is less than 1% and the conversion ranges between the winner and loser don’t overlap.
- Run each test for at least one business cycle (those curves in your Google Analytics visitor stats) even if confidence level is achieved sooner. Typically a week, ideally two.
- Tests that you run during massive ad campaigns or Christmas are skewed, results might not carry over once those pass.
- Test one part or segment at a time. If you have concurrent tests on multiple pages, your results might get skewed.
Read these 2 great resources on split testing:
#10. Learn to prioritize
You can’t test everything at once, and traffic is precious. So you need to know how to prioritize your tests.
There are 2 great methods for prioritizing tests.
1. Use the PXL Prioritization Framework. It’s a binary model that asks you to bring data to the table. We developed it after many years of running tests for clients and on our own sites. Check it out here.
2. The Ring Model by Craig Sullivan. Ring model is a way to look at the ‘layers’ or ‘levels’ reached. This works for many (but not all) websites. It focuses on depth of engagement, not pages viewed.
It’s similar to a funnel as it helps you see the key loss steps. The main point is to help you see the ‘big picture’ involved.
The Ring Model helps you see where flow is “stuck”—where the traffic is not flowing down to the next level. The main benefit here is that it helps you see which layer of your website needs the most help.
What you do here is you measure unique pageviews per layer / step. If you had 300k product page views, 5k adds and 1k checkouts where would your problem be? Cart adds! If you had 300k product page views, 100k adds and 1k checkouts your problem is in a different place!
From now on it’s going to be more and more about data and analytics—measured success, no more bullshit. Get with it, or stay behind.