Beginner’s Guide to Marketing Funnel Analysis
UPDATE: Download a free PDF version of our Marketing Funnel Analysis guide!
“Tell me what’s happening with marketing.”
As a junior marketing executive I got cold-called by our CEO late at night with that deceptively simple sounding question.
“Ummm, well Google, the ads, they’re good and we’re still getting a lot of users from Organic Search. We, uh, wanted to test Facebook ads …they seem to be working. I mean it’s driving a lot of clicks, but our revenue per user is down and I think it might be the Facebook traffic.”
Pro tip: if you want to keep your job, don’t get in the habit of giving your CEO vague answers.
A better answer would have been:
“Google AdWords and Organic Search drive 80% of our new subscribers, but growth is flat year-on-year so we’re exploring Facebook ads as a new channel. We’ve spent $600 so far, and we’re acquiring new users at half our $2.50 CPA target. That said, these users are also buying half as much, so they net out at the same ROI.”
Awesome. Now how do you do that?
How do you look at data and figure any of this out? How do you know what’s important?
Well it starts with a good tracking and tagging setup, but assuming that part is already done, how do you make sense of a big mass of user data?
Need an in-depth analysis of your marketing funnel? Ladder can help:
The Analysis Process
I created the following graphics for a talk I gave at Grand Central Tech last year. They’re much simpler (and prettier) than what you’ll see when you log into Google Analytics, so hopefully they’ll be less scary and easier to understand.
Let’s say you have a mobile app with 10,000 users.
Not all users should be treated equally; some will be new to us, others returning, and a smaller amount will potentially have bought something. Let’s split them up:
What we have now is a basic marketing funnel. But what does it tell us?
Well if it looked like the one above instead, quite a bit: we’re acquiring a lot of new users, but not many are sticking around or becoming customers.
Knowing this, we’d now want to focus on improving our conversion and retention. This is an important step: just by splitting the data by stage in the buying cycle, we are already able to use the data to define our strategy.
If we had the exact opposite situation, a large customer base but very few new users, we’d be much happier. This is a sign we’ve got a strong product offering.
However this information would beg the question: with such a strong product, why are we not investing more in growth? We can surely afford to increase our marketing budget to bring in new users to convert?
But before we start investing, we want to be truly sure we’ll make our money back. Let’s break our customer base down a little further.
This is great; the vast majority are buying multiple times. This is a marketers’ dream.
This however, is more of a nightmare; they buy once and they never come back for more. This insight might make us revisit our product to fix any issues before we double down on marketing.
Of course, for most businesses, there’s a step before buying something – signing up. With 1,000 people’s email addresses on file who haven’t bought anything, maybe we should build a CRM campaign to drive more sales?
At this point CRM is a no-brainer; drop everything else and figure out how to cash in on your 4,000-strong email list.
So now we’ve sliced the data horizontally (aka a marketing funnel). What are some other ways to look at the data?
Slicing the data vertically by marketing channel shows us right away just how dependent on Facebook ads we are for acquisition: it drives over 70% of new users. It also tells us how relatively unimportant Google Adwords or the App Store are in comparison.
Now if in your experience the App Store drives a larger percentage of users than this normally, you might want to focus on this channel. If 7% of new users is normal for App Store contribution, you are safe in the knowledge that you’re doing ok there.
What, did you expect all channels to convert equally? That’s almost never the case. In this case, though the App Store only drives half the new users of Twitter, the conversion rate to returning user is so high, it ends up on par. Similarly though Facebook is still the king, there is a significant drop-off which indicates a lower quality of user.
Going deeper down to the end of the funnel is truly revealing. Despite driving the most new users, the low conversion rate means Facebook barely drives more customers than Twitter and the App Store. Google Adwords isn’t driving any customers.
Of course at this stage it’s important to introduce marketing spend. If you’re dumping thousands of dollars into Facebook, you might be losing money on those 354 customers. If Twitter is relatively cheap, you might want to shift some of the budget that way.
In my experience it’s also not unusual to see one channel to work exponentially better than all others. That’s why it’s so important to conduct this sort of full funnel analysis – so you know what to shut off and where to redirect the budget.
As well as segmenting by channel, you should also be segmenting by time. If each ‘funnel’ above represents a month’s worth of data, this business looks like it’ll be dead in a few months.
This one’s doing quite alright; not only is it trending upwards, but all funnel stages are growing, which is great to see and means this growth is likely to be sustainable over a longer time period.
This is even better; as time goes on we’re growing our customer base exponentially; this is a strong sign that we can start to increase our investment in marketing.
…but it means nothing if we’re nowhere close to hitting our goal. Missing our target could mean inability to raise an investment round, staff downsizing or even bankruptcy.
In this scenario, we’d be pretty happy; it looks like we’ll comfortably hit the goal on time based on our projections…
Be warned: past performance isn’t always a reliable indicator of what’ll happen next. Maybe we took the foot off the gas because we got too comfortable? Maybe a key competitor started poaching customers or we had a huge quality issue with the product?
This is a reminder that data isn’t magic. The chart doesn’t go up and to the right because you want it to. It’s just an abstraction of the activity of real people and it’s important understand factors outside what you can measure in your analytics platform.
It’s also useful to remember that we’re looking at simplified charts here; real data is messy and the patterns are harder to find. Don’t be disheartened when you don’t get a smooth graph; most of the real data I’ve seen looks something like the chart above. It’s as much an art as it is a science to interpret data sometimes.
Despite all the irregularity we observe in data, we still have to make our best guesses at what’s happening or we can’t run the business. See the example is above, where I’ve mapped out what the ‘Pirate metrics’ (the acronym spells AAARRR) might look like for a business.
Once you have the conversion rates and the final revenue amount your customer pays you on average, working out the ‘Cost’ column is easy. Just work backwards by multiplying the Cost from a previous stage by the conversion rate. I.e. if a signup is worth $13.33 to us and 5% of visitors signup, a visitor is worth 67 cents (13.33 x 0.05 = 0.6665).
In the real world, shortfalls in the way tracking solutions work means it’s almost impossible to get these funnel numbers accurate. However if you don’t at least have a good assumption in place (or as close to real data as possible), you’ll be plagued by analysis paralysis.
If you didn’t know how much an ‘impression’ was worth to you, how would you decide whether to run a brand awareness campaign? If you aren’t sure how much you can afford to pay per click, how would you know if a new channel is viable? If you don’t know your conversion rate between each stage, how do you know which part is weakest in order to focus your efforts?
What data surprises you? What data doesn’t fit with what you expected? It’s only by continually segmenting and interpreting the data that you’ll build a good mental model of what to expect. Once you have that model, you’ll be able to make good decisions without even having to run the numbers.
Making good decisions based on a mental model built with solid assumptions and backed by in-depth data analysis is really what growth strategy is all about.
So what did we just do? Let’s dive into a little more detail on the principles behind this process, so that you might replicate it for yourself.
Download a free PDF version of our Marketing Funnel Analysis guide!
Insight, Context, Action.
At Ladder, when conducting analysis, we developed a framework called ICA, or Insight, Context, Action. It works like this:
- Insight: “My Google ads have a cost per click of 80 cents”
- Context: “On average, clicks are worth over $2.50 to us”
- Action: “It’s time to push up our spend on Google ads”
Insight is worth nothing without context.
If I didn’t know how much a click was worth to us, knowing my Google ads CPC means nothing to me. Sure, it’s nice to know, but it doesn’t help me take action.
Context is useless if it doesn’t drive action.
To quote the SEO agency Distilled, “It’s NOT our job to deliver reports. It’s our job to effect change”. If we’re not taking action to improve our marketing campaigns then any analysis we just did was largely wasted effort.
One great source of context is Benchmarking.
Knowing that the global average Facebook CPC is $0.48 tells you whether your campaign is killing it or has room to improve.
Knowing that exit-intent email capture popups convert 1.28% of visitors, helps you plan how quickly your email list will grow with your website traffic.
Knowing that over 57% of marketing experiments fail helps you set realistic expectations on the success you’ll see from your testing plan.
You should take these industry benchmarks with a pinch of salt; these are just averages of many businesses that are completely unlike your business.
I’ve seen ROI positive campaigns with $30 CPCs and loss-making campaigns getting under 12 cents a click. It all depends on your business model, and you may have strengths in one part of the funnel that compensate for weaknesses in another part.
It’s up to you to figure out if the benchmark is relevant.
The ability to set good benchmarks is one of the major advantages a senior marketer has over a junior marketer. They’ve seen it all before, so they know what’s possible, and what’s probable.
At Ladder we’re trying to level that playing field: check out our tactic playbook where we share over 480 proven growth tactics, so you know what’s possible, with expected ROI for each one so you know what outcome is probable.
Of course the very best benchmark is against yourself: that’s where segmentation comes in.
Segmenting the data just means you’re splitting it by one or more variables and seeing how each ‘segment’ compares.
You could split your data by marketing channel to see which are working and which aren’t. That then tells you where to put your budget.
Splitting the data by time is another effective segment. Is your marketing doing better this year vs last year? This month vs last month? This week vs. last week? Knowing this context can tell you whether to sit back and relax or to go into panic mode.
Other popular segments are audience (are teachers or stay at home moms buying more?), location (do New Yorkers spend more than Texans?), product (is the bump in sales due to rain jackets or hiking shoes?) and employee (is Sally outselling Jane?).
The cleanest form of segmentation is A/B Testing. You’re running one variation vs. another at the same time to see which works. Is this new ad copy performing better than your existing ad copy? Is this new landing page performing better than your existing landing page?
As long as you’re always beating your old campaigns, you’re growing.
And ultimately, that’s your job.
Not to pull reports or to produce shiny graphs, but to drive growth.
At Ladder, we build software and offer services to help high-potential businesses accelerate their growth. Is that something you need?