Category Archives

8 Articles

Posted by Jeromy Sonne on

Your Metrics Obsession Is Killing Your Marketing.

I’ve always been a stickler about metrics. I’m actually known for a phrase that has led some people to dub me as “inflammatory” or an “asshole” that has to do with my love of marketing metrics.

“Your opinion on marketing doesn’t matter.”

Now of course that’s nothing but a truism designed to shock people into changing their mindset about how marketing works. For the vast majority of people data driven marketing is a new thing. For even more people interpreting the results correctly is even more difficult. Let me explain.

You’re reading this on Medium. Likely you came here either from one of our social properties, reddit, Hacker News, or some other similar technologically savvy community. You likely have made the jump from 0 to 1 when it comes to thinking about marketing and data. That is, you understand that modern marketing can be measured for it’s effectiveness, it can be tested and iterated on, and marketing can be tied to revenue.

You are probably just dangerous enough to kill your own marketing program.

Now this isn’t intended for me to be a jerk, although as I re-read this I’m starting to see why I have that reputation. Rather it’s to point out that knowing enough to be dangerous isn’t the same as truly knowing. When you know the lingo and jargon but don’t truly understand what it means, you can often do more harm than good.

A great example. I hear people, especially in the tech community, running around saying things like “value” and “conversions” as goals of their marketing programs all the time. Yet very few people understand what either of those things mean, nor how they’re defining it.

I hear a word like value, and that doesn’t mean anything particular to me. It’s so vague that, to me, it could mean any of the following:

  • Return on ad spend
  • Return on investment
  • Shiny things people like
  • Things your nephew who is good with computers likes
  • Awareness
  • Brand Recall
  • Market Penetration
  • Authority
  • etc. etc.

The truth is value is defined by each individual largely as what’s important to them. They substitute their own biases for objective reality and expect, either maliciously or in a more benign sense, for everyone to perceive and anticipate what they want. What’s dangerous about this is that data driven marketing is not a personal experience of marketing, it’s a thing. More so it’s a thing that you don’t get to decide, your customers do.

Conversions is another one of these. People shout from the rooftops they need conversions, but conversions to what? Conversion to get an email to a newsletter? To a customer? To visit a specific page or perform a specific action? This is a term that has great value (ha!) that has been stomped into the ground by people posturing and trying to say the right things. When you think about conversion as literally just x to y in a granular sense, you can work on your funnel piece by piece rather than running around like a chicken with its head cut off.

This also manifests in other ways like optimizing for a specific metric like click through rate, or cost per click etc. People start focusing down on these things like they’re a silver bullet that will fix everything. Nah.

Metrics are only valuable in the context of other metrics. Having a dirt cheap cpc isn’t worth anything to me if it’s an irrelevant and un-engaged audience. A great ctr is meaningless if the people clicking through aren’t my audience. A form completion conversion is wasted if the potential customer isn’t qualified. In fact I can envision many scenarios where these metrics could be in the metaphorical toilet where I would be happy. It’s all about how they work together, not alone.

I’ve seen it before, and I’ll see it again, otherwise smart people getting stuck on specific points for seemingly random or emotions based reasons. I say all of this not to be a jerk, but to help you. Too many companies and teams that I see have brilliant insights in front of them that could make them successful but they’re too blind or self interested to seize very real opportunities. Don’t be the team that focuses down on specific things you feel are important. Be the team that looks at all the data and knows what the market is saying to them.

The point of this long winded rant is basically this. It’s awesome there’s a lot more people thinking about data than there were even 6 months ago. If you’re thinking about data that’s great. Learn more. Try stuff. Ask lots of questions. Don’t get tunnel vision, and understand that you know enough to know there’s a lot that you don’t understand. That’s okay. I wake up every day learning about things I’ve never heard of as a marketer. If you don’t you’re stagnating, and in this industry stagnation is akin to death. So be confident and bold, and always keep learning and testing. Just don’t become arrogant and reliant on your incomplete knowledge set.

Your Obsession With Metrics Is Killing Your Marketing. was originally published in Observed Reflections.

Posted by Karl Taylor on

“We’ve Tapped Out The Universe”

Advice For When User Growth Starts To Slip

It’s an all too common occurrence.

A marketing strategy that had been working, shows signs of slowing down.

Stakeholders start to get nervous.

A frantic team works to figure out what can be corrected, when and where.

Situations like the one I’m describing require more than increased output, but all too often, this is the solution I see teams applying.

Left unchecked, a situation like this one can quickly get out of control.

I’ve written a little bit in the past about a handful of ways you can identify what you should test, as well as a handful of ways you can use information like geographic breakdowns to better explore anomalies in click data, but getting value from information like this presupposes that you haven’t encountered a situation that to be frank, teams often do.

Sometimes what you’re doing stops working.

There are a lot of different ways that you can take a step back from the situation you find yourself in. For some, there is a very real temptation to start identifying each way that a program has failed to meet an objective.

The trouble with this approach is that things don’t really happen in a vacuum. When your ad begins to saturate, for example, you might only notice when the cost per click begins to rise. Information like cost per click proves retroactively useful, but it doesn’t necessarily shed any light into a problem that’s developing. As an example, it is quite possible that you could have detected a problem before it developed by observing ad frequency or response rate.

While it may be true that large portions of a marketing program may need to change, it is unlikely for example, that lists of customers are truly worthless. Design assets have a much longer shelf-life than many businesses allow them to enjoy. Recognizing the value in the efforts is part of the path to getting back on the right track.

More often than not, the reaction to proclaim “everything is broken,” is a reaction stemming from a frustration that has a good reason for existing. Instead of continuing to apply pressure to the situation, it is often a good idea to step back and take a look at the work that is actually being performed.

Many times this can reveal powerful opportunities for growth that had previously gone neglected. For example, many platforms have similar sizing requirements for assets. It is likely possible to begin experimenting in a new direction with minimal startup cost. This doesn’t mean, of course, that you can use the same strategies on every platform, but if you have good performing copy from a blog post, there’s no reason not to see what happens when you place it somewhere else. You can always go back and improve on a promising development.

That’s why when you find yourself in a situation like this, the only way out is to make a left turn. Ask yourself why you’re making what you’re making. Is there a tactic that could be more effective at reaching that goal? Giving yourself the space to take a broader approach can be just the change you need to get through.

“We’ve Tapped Out The Universe” was originally published in Observed Reflections.

Posted by Sonne Taylor on

“But How Long Will This Take”

The Time It Takes For Your “Marketing To Work.”

I’m going to let you in on a little secret.

Imagine you have a mobile app that gets 100 downloads a day and you’re trying to decide if you should invest in marketing.

For every 100 people who download your app, 5 become a paying customer. You’re averaging 5 new customers a day. For the sake of easy to follow numbers, let’s assume that each new customer is worth 10$ to your business — after costs and the App Store’s cut are taken out. You’re making 50$ a day.

Now you may say to yourself, self how can I make this 50$ work for me in the form advertising.

You draw up a series of ten ads. You build an audience of 500,000 people. You give each ad a budget of $5 dollars*. This means that each ad is set to reach an estimated audience of 5,000 people.

Between 2–5% of the people who see an ad, click. We’ll use 2.5% as the average. Of the people who click, 20% of the people download the app.

That means your ads that reached 50,000 people also generated 1,250 app downloads. For every 100 downloads, you generate 5 paying customers. That means your ads generated 12.5 paying customers. (We’ll say 12.)

If each paying customer was worth 10$ to your business, we could say that your 50$ of ads generated 120$ of profit.

You more than doubled your money.

But that still doesn’t answer the question we set out with: the question of how long it would take to reach your goals.

The trouble is, that question is nearly impossible to answer without some insight as to what your particular goal may be.

What do I mean by this?

If you take a second walk through the example you’ll notice my asterisk. It’s placed there because in creating that thought experiment, I made a very very generous assumption about what astute advertisers might have noticed as CPM.

The truth is, there are very few places you can advertise where $1 of ad spend could reach 1,000 people in your target audience.

It’s also unlikely that a new app would be generating a conversion rate of 5% out of the gate — especially at the price point you’d need to walk away with 10$ per sale.

There are a lot of levers to fiddle with.

But when it comes to calculating how long your ads need to serve to reach your goal, the levers you need to be most concerned with are the ones relating to how many people see your ad, what percentage of them take action and what percentage of those complete the action.

If you have that information, you can figure out how many people you need to reach and at what response rate to generate the number people taking action you’re looking for.

From there, you can figure out how long you’ll need that action to occur in order to hit your goal.

If you need to move at a faster pace, you can increase the number of people you reach at the top of the funnel. If you need to stretch your dollar, you can look for ways to increase conversion rate or slow down the rate at which you reach your audience.

While you won’t ever get numbers quite as clean as you could in our little example above, you will find this a very helpful rule of thumb for calculating just how long you’ll need to advertise.

“But How Long Will This Take” was originally published in .

Posted by Karl Taylor on

What Do We Know What To Test?

Some of the more curious questions we get from marketers involve testing.

I’ve found this strange, in part, because I’m not sure I’ve ever once wondered why anyone would need help figuring out what experiment to run next.

A quick glance through Google shows me that there are quite a few resources like Gabriel Weinberg’s32 Fast And Cheap Marketing Tests You Should Consider Running.” There are even collections of results like the delightful “I Spent $30,000” (Justin Brooke) and this great classic on meaningful tests.

I’m not sure I’ve noticed too many places where anyone explores how you decide what marketing experimentation means for your organization. The truth is that at the end of the day, you’ve got to have those goals in place before you can effectively conduct any sort of marketing experimentation.

It may be a cliche, but if you’re letting those goals guide your work, they’re going to guide your research, too. Try and be cognizant of how this may affect what you’re looking at. Sometimes it might mean you need to look at a new technique or tool. In other situations, it may mean that you need to reevaluate your approach.

Next, you’ve got to find the big picture questions you need to have answers to. Perhaps you are testing a new design or experience and are looking to collect feedback. You may need to answer a question about product cyclicality. You might want to know who your customers are.

For example, if I really knew the customers of a client, I would have a list of customer profiles and email addresses. I could talk to each person individually, or I could make a list and buy TV ads and send postcards out until I heard from each one. I could also place ads on websites and in smartphone apps my customers frequented.

Regardless, without a list like that, I don’t have a real 100% answer to “who is your customer,” and that’s why it’s a great example of a big picture question.

Once you’ve got your questions written down, you should figure out how important each one is.

You should already have some idea of how important each question as you made this decision once before (when planning out your company’s purpose, priorities and values.)

This should be enough structure to fight even the most stubborn cases of mind-block. If not, this is a good time to explore the wilds of delightful ideas, hacks and tips that get added to the internet every day.

Pick the strategies and tactics that will lead you towards an answer to your question. If you need to prioritize further you have a few options. Two of the most straightforward: you can sort your list of tactics by resources required (time or money involved are two common methods) or you can try and figure out the most efficient way to finish the full set.

Once you’ve assembled a collection of tactics to help answer your questions sorted by priority, you should take a little time to make sure you finish your work.

This is particularly important as you are likely to discover new questions while you are conducting your research.

One of my favorite ideas was recently very well explored by Srinivas Rao who shares a tip for using a calendar to conduct research in “Why Calendars Are More Effective Than To-Do Lists.

You can use this approach to make sure you keep everything moving ahead on schedule.

This won’t always work, and there really are some problems out there that are going to take a lot more to solve than this sort of simple arrangement can allow for. That doesn’t mean it isn’t worthwhile.

I’ve found that over the last few weeks a similar approach has reduced interdepartmental friction.

One problem I’ve noticed on a lot of teams is that while the folks who work on the marketing may understand what is going on, it’s very difficult to communicate that information to individuals who haven’t had the chance to look at it.

There’s just something that gets better when you’ve put together an experimentation that everyone can understand.

Working from the foundation of shared principles means you don’t have to worry about buy-in. When the whole team knows what the experiment is at the outset, they know what sort of result to expect at the conclusion.

What Do We Know What To Test? was originally published in Observed Reflections.

Posted by Sonne Taylor on

Persistent Selling

There’s a troubling trend I see in startup marketing plans. It’s almost solely focused on user acquisition, with absolutely 0 plan to activate, retain, fight churn, or leverage their customers for word of mouth. It’s a sad state of affairs.

It’s essentially the marketing equivalent of thinking “That Iceberg isn’t so big”, ignoring that 90% of the importance in marketing comes after you acquire a user.

I don’t think it’s any one person’s fault. I think that there’s a few startup “memes” that get traded around enough that it’s finally done real lasting damage to marketers ability to do their jobs effectively.

“Move fast and break things” aka let’s torch brand equity.

The obsession over top line numbers to appease investors is incredibly harmful. Who cares if you’re getting $0.10 cost per installs if your customer lifetime value is $0.05 or you have 99% churn.

The general need to be “killing it” all the time that’s translated only into easily understood top-line metrics rather than creating meaningful, long term relationships with your customers.

Frankly, just plain not giving a shit and being lazy.

This notion that you should ignore your competition.

Want to get in on a little secret? Your competition is poaching your customers constantly. I do it all day every day. Nice little Twitter audience you’ve built up. Would be a shame if someone scraped it, made a custom audience, and messaged your customers based on your biggest weaknesses. Your customer churn is my gain, and there’s dozens of me going after you every single day. The only way to win is to be constantly selling to your customers and continuing to grow the relationship. In the world of business, and specifically marketing, you are never “done”. You grow, or you die.

Persistent Selling was originally published in Observed Reflections.

Posted by Sonne Taylor on

The Problem With “Widget, 5$.”

A few years back, the advertising world was caught up in a debate over the relative values of “Rational” and “Emotional” messaging. Emergent digital targeting data and performance analytics made quick work of the conversation…

…or you’d think so, but that would make too much sense.

There’s a misconception that’s still all too common amongst teams that aren’t up to spec on performance data. This idea is so toxic in part because it means many believe that if they don’t berate the customer, the customer won’t know you’re selling a widget and they won’t know what to do with the information you’re sharing with them.

Don’t get me wrong: strong calls to action are important, they drive results. They promote cohesion. They keep things moving.

But teams who think this way are at a serious disadvantage when it comes to future proofing their marketing program.

For a long time, cross-device performance was a pain to track.

A company might know that you visited a website on a desktop computer. They might also know that later a user made a purchase from a smartphone.

Historically, linking events like those presented a fairly significant challenge.

It’s easy to understand that your messaging isn’t being consumed by your customers in a vacuum, but it’s a lot more difficult to take a leap of faith and trust that a % of the people who see your ad on Facebook will launch a new tab and start looking up additional information about your product or service on Google.

It’s easy to intuit that when someone engages with a Snap that features a product, they’ll have an association with that product when they see it again in a display advertisement.

One very slight adjustment to the way you look at the world can be enough to change everything. The truth is, that cross-device is still out of the budget for a lot of digital marketers. I don’t think that’s a good enough reason to stick with the old world, though.

At the end of the day, you’re very likely a native user of many of the platforms that your team uses to promote itself.

You don’t have to grasp for straws to figure out where the balance between “savvy” and “scammy” is, you’ve got to roll your sleeves and start looking for it.

The Problem With “Widget, 5$.” was originally published in Observed Reflections.

Posted by Sonne Taylor on

Quick Suggestion For A Poor Team Needing An Early Analytics Stack

I think all too often, startups put off putting in basic tracking capabilities until they feel like they can afford a solution on the level of Adobe Marketing Cloud — or in the medium term HubSpot.

I think if I were just getting started on a project this is the flow I’d use. (It’s actually the flow we’re using on our team’s branded properties right now.)

  1. Install Segment.

I really shouldn’t be so shameless, but if I were going to deploy a tag on an early stage team (say an early website) I wouldn’t want to have to revisit updating my site each time I discovered a new capability I wanted to play with.

You’re likely going to be doing a lot of that in the beginning. It’s a good idea to get started in a way that you’re going to find useful as you go on.

(You could also definitely wrestle with Google tag manager. 9 Clouds has a solid primer going on here.)

2. Integrate The Properties You’re Currently Using

It’s likely that you’ve got a Facebook page, and probably a Twitter account. You should take the opportunity to get signed up at and follow their directions to setup a pixel. (I like this guide from Pavel Matvienko)

You’ll find that the fine folks at Segment have a number of delightful guides that should work you through updating the settings for each integration — which in the short term will save you from having to make additional changes to your website.

You don’t have to go crazy here, because you’ll find you’re in a plan that’s priced by MTUs, which if you’re finding this advice useful, you’re a ways off from having a big number of.

3. Website? Google Analytics. App? Mixpanel.

In both cases, you should settle pretty near to the free tier — it takes a while to outgrow.

If you want to drill a little deeper, here’s a good introduction to Segment from Francois de Fitte.

Quick Suggestion For A Poor Team Needing An Early Analytics Stack was originally published in Observed Reflections.

Posted by Sonne Taylor on

What To Do When The Digital Audience Information Doesn’t Make Any Sense

Even This Looks Better Than My Versioning Flow.

I really need a better versioning best practice.

The system we use now can be best characterized as an artistic interpretation of Google Drive. It’s a mess of columns, files and rows. It’s held together with bailing wire. It’s held together with willpower.

In short, it’s a mess.

But it’s my mess, and I daresay I’ve even grown fond of the time I’ve spent revisiting old work and cleaning up headers.

I was doing just that a few nights ago and stumbled into something that reminded me of a story I haven’t told many people before.

It was 2013 and I was working with an enthusiastic small team to validate a market opportunity many well-reasoned investors would have considered “overly niche.”

They were looking for a cross section of student that didn’t really have any unique media consumption habits.

Like any good marketer, I took a step back and looked at what information I knew that I did have.

I knew a lot about the types of people who bought their product. I knew their favorite TV show (Grey’s Anatomy,) I knew their favorite shoe (those foam flip-flops you get at the dollar store.) I even knew their favorite cookie (chocolate chip, obviously.)

What I didn’t know, was how I could help this team reach their audience. There just weren’t the resources to target their ads to every person who liked those three things. Lots of people like those three things! I like those three things!

Competitors in their space had resorted to an all-out ground based strategy.

Each competitor would set up a booth at each campus at each school each semester. Their field representatives would speak not only to the students, but the parents. It didn’t matter that we had better content. It didn’t matter that our option was cheaper and easier to use.

Once they had succeeded in convincing the parents of a new college student to buy something, that college student was getting that thing.

In the face of those kinds of odds, mounting the resources necessary to compete 1:1 seemed very unlikely — but targeting each person who could possibly be in the universe wasn’t going to get the job done, either.

I decided that if I could find a way to narrow the number of people we were looking at online in the first place, I’d have a much easier time measuring what I needed to get the team’s message in front of the right people.

At the time Facebook had just launched location based targeting. You could target by city, country or zip code.

Targeting each city wouldn’t work because I’d have the same problem I had before. I didn’t know enough about the people who might find the product valuable to figure out how to reach them.

Something occurred to me.

Facebook wasn’t offering a bulk import feature at the time, so if you wanted to do anything except for target cities, you were going to have to manually identify and upload zip codes one a time.

That’s just the kind of crazy scrappy super inefficient thing a small team can do that a big entrenched player won’t even dream about doing. It was exactly what I was looking for.

So I and a fellow contractor pitched a special project. We started assembling a list of the zip codes of campus housing around schools where we might find the students we were looking for.

The Real Winner? Publically Available Data That’s Collected By Experts

I could tell you the story of the time I spent out of my own pocket putting each one of these codes in — a job I really should have subcontracted — but that’s not what this post is about.

When I added in a bit of real world data — data I knew was better than the data the media platform I was buying on was giving me, I didn’t have to worry about an overly niche market anymore.

I didn’t have to worry, because by combining these two sources of information I had a better definition of what the market really was.

It’s easy to think that because information that has historically been used to put people into boxes exists, the question is about whether or not you should use it. I think that because of this, many young teams try to avoid using the information they have at their fingertips because they’re worried about becoming out of touch — just like their competition.

Metrics like these aren’t why your competitors are out of touch. Metrics like these are why your competitors can still do business despite being so out of touch.

Having a better understanding of who your ideal customer is than your competitors do changes the way you think about the people who buy your product. Paying attention to these indicators gives you a language to plan around what your customers tell you they need — not what you think they want.

That’s why it’s so important that you learn not to be overwhelmed when you stumble into some information that makes you rethink your marketing strategy. When you see that only 2/100 of your email list sign ups follow on you on Twitter, when you learn that your Snapchat universe is reaching 2,000 people you don’t have any information on, when you wonder if buying an ad on a new podcast is the right idea…these are the times when you should be excited.

These are times when you should look for answers from unconventional sources. These are the times when you should consider unconventional applications of conventional sources. These are the times where it matters that you try your best — because you only stumble into these problems when you’re actually doing something that no one has ever done before.

What To Do When The Digital Audience Information Doesn’t Make Any Sense was originally published in Observed Reflections.