At Ladder, we believe in transparency as a smart way of doing business. In that vein, we’re open-sourcing our 2017 marketing plan for growing Ladder. Starting next week, we’ll begin releasing bits and pieces of the growth auditing process we’re running on Ladder’s current marketing process every week, culminating in the release of our full 2017 marketing plan in January.
As a precursor to that, we’re updating and re-publishing this post (which originally appeared on Chatbots Magazine) to give you an idea of the marketing plan we’re releasing.
Side note: The plan below was made with publicly available information and is therefore limited by the dataset on offer. For the Ladder Marketing Plan, we’ll be open-sourcing our analytics so you can see exactly what we’re doing, how we’re doing it, and how it impacts our bottom line.
Too many times I get asked how to write a marketing plan. And as the founder of a marketing agency, I know all too well how badly things go without one.
Most people have no idea how to write a marketing plan.
Worse, when you Google “Marketing Plan” you end up with a list of theoretical, non-actionable templates that won’t get you anywhere.
Because showing is better than telling, I decided to put together my own style of actionable marketing plan that I can point people to as an example. To make things interesting, I chose a real life startup, And Chill, to focus on.
Although I know the Founder, Jake, and I use the product (it’s awesome!), he had NO IDEA that I would be doing this and provided no special information. Everything below was pulled together from free tools and publicly available information.
This is a full marketing plan — we’re going deep on tracking and analytics setup, SEO, paid advertising, CRM, landing page optimization, virality and even branding. This is what you should expect your marketing lead to deliver, at varying levels of quality based on experience. It should be even more in-depth if you have existing campaigns that need auditing. But don’t expect this exercise to be easy — it took me about 40 hours to pull this all together and another 10 hours to write it all up!
The real benefit of this exercise is knowledge — through the course of researching and planning And Chill’s marketing, I got to know the movie recommendation space intimately. I’m now familiar with the major players, and can talk confidently about what seems to work (or doesn’t), and exactly how And Chill is better than the alternatives. This is all knowledge that a senior leader in a business HAS to have to remain credible and effective. Who would you choose to execute your marketing? Someone who did this exercise or someone who showed up on day 1 expecting to wing it?
Close to 100% of the companies that come to me for advice have wasted a HUGE amount of time and/or money by choosing the wrong marketing partner, making the wrong hire, budgeting incorrectly or focusing on the wrong activities. With a small investment up front in a plan, you’re more likely to spot your error before it happens. Just one avoided mistake can save you tens of thousands of dollars or hours. One smart decision that helps you grow faster could add millions of dollars onto your company’s valuation.
I hope this helps you with your own plan. If you get stuck, or want something similar for your startup, tweet at me: @2michaeltaylor
1) About And Chill
2) Market Research
3) Marketing Plan
4) PPC Campaigns
5) Budget Forecasting
Note: Missing from this list are a whole slew of internal audits, from a PPC audit to an analytics audit and more, as that data isn’t made publicly available by And Chill.
For the Ladder 2017 Marketing Plan, we’ll be publishing our own internal audits with actual data, all over the next month.
Company: And Chill
Founder: Jake Cohn (https://www.linkedin.com/in/jacobmcohn)
Tagline: We’ll give you spot on movie recommendations you’ll love. No headaches. No endless browsing.
Summary:
Movies movies everywhere, and not a thing to watch! We’ve all been there, endlessly trawling Netflix searching for the improbable — an instant favorite that you haven’t seen yet. And Chill solves that problem. Even better, it does so in a friendly, highly personalized manner, through a chatbot you can access just by texting a phone number.
And Chill has a unique opportunity. Advancements in AI and a big ecosystem push from the major messaging platforms has investors excited about chatbots. Netflix has shown precedent that they’d pay to improve recommendations. Online content is proliferating as the competition between the major players heats up and they invest their war chests in original content, which will only make the ‘what to watch’ problem worse.
Others have sensed this opportunity. There are so many Netflix-focused apps, Product Hunt has a whole category. Many of these tools aren’t fit for purpose and soon become defunct. Publishers are getting into the game, attracting more eyeballs to ads. Most of this is poorly done clickbait — an after thought from writers running the pop culture beat. But the demand from the public is there — waiting for the right product to come along. I believe And Chill is good enough product to win, so how make sure it does?
Another document that’s useful to share with designers and copywriters is the brand style-guide. This is your definition of your brand — how things should look, feel and sound. You can start relatively simple (as I have here), and build on it as time goes.
Elevator Pitch:
For (target customer)
Who (statement of need or opportunity),
(Product name) is a (product category)
That (statement of key benefit).
Unlike (competing alternative)
(Product name)(statement of primary differentiation).
For Netflix Subscribers who waste time endlessly browsing, And Chill is a movie recommendation service that makes truly personalized movie recommendations.
Unlike Netflix’s recommendations, And Chill is a 2 way conversation — chat to our AI bot today and find movies you’d never find yourself.
Colors:
Fonts:
Tone of Voice:
Examples:
The final piece, SOCO and SOCA are the two things you should be armed with in any press interview. They dictate the one message a journalist should be left with, and the one thing you should avoid being dragged into at all costs (even if it means talking like a politician).
SOCO (Single Overriding Communications Objective):
Conversational AI is the future of recommendations.
SOCA (Single Overriding Communications Avoidance):
Netflix’s big data is superior to conversational AI.
Contents:
a) Persona Development
b) Competitor Research
c) Proposition Mapping
d) Value Propositions
e) Keyword Research
f) Technical SEO
If we want to solve a problem, it helps to know a little about the people we’re trying to solve it for. The goal of researching personas is to understand how your potential users think and act. That way you’ll know the appropriate time, place to reach them, with a message they’ll actually be receptive to. Most people wing this part, to be honest, and assume they know who the customer is. They may be right, but personas form the bedrock for every decision you make in marketing — wouldn’t it help to spend at least a couple hours digging through data to make sure?
Netflix Research
Sources: (1), (2), (3), (4), (5)
I picked all these facts up with a few minutes Googling. So what do they tell us? Well it tells the market is huge — 81m subscribers (that’s not counting the other streaming services And Chill could cater for). It’ll only get bigger as the millennials grow up and age out the baby boomers who aren’t using Netflix. It’s also a big part of people’s lives — 70% binge watch shows and 7.7 hours of content watched per week. It also gives us a figure we could use in our marketing — with 5,532 titles how will you possibly choose from all of them? Finally it gives us a sense of urgency — if the typical user only looks at 10–20 rows and gives the task only 60–90 seconds, we know And Chill’s recommendations have to fit within that window.
Next we’ll look at Facebook data. Not everyone knows this, but if you set up a Facebook ads account, there’s a tool called ‘audience insights’ where you can filter for demographic and likes / interest data to get an idea of your audience size. Here’s what I got from playing around for a little while:
Netflix Fans:
Does that quell any assumptions you had? For me I didn’t expect Netflix fans to be less likely to be married. I also expected much more on iPhone than Android. It’s pretty stark how dominant mobile is in this space — And Chill is doing the right thing by starting mobile first. The household income one was a surprise — I saw video streaming as something my more privileged friends were into. You’ll probably have different reactions to this… or maybe you guessed all of this exactly right (congrats!). But as you go through the rest of the data in this plan, keep track of how many times your assumptions were wrong — it’ll give you a good sense of why this is a valuable exercise.
Next we’re going to drill down into specific personas. Now this is more of an art than a science, but to arrive at these groups I followed this procedure:
Doing this I arrived at 3 pretty interesting segments, all of which represent a substantial (but niche enough) ‘type’ of Netflix user. To give you a feel for the process, I’ll run through the first one:
Millennial Couples // 375k
GenX with Kids // 275k
Blue Collar Single Men // 375k
Of course there are a number of different personas here, but I selected the above showed a significant bias in the data vs the general population i.e. Netflix users were much more likely to have these traits than average. This is part science, part art, so play around with different segments until something in the data jumps out at you.
Usually those who create personas without data end up with a carbon copy of themselves (or close friends). To see the value of starting with data instead, try filtering for your own persona to see how many people there actually are that have similar demographics to you. Chances are, even with a pretty broad definition, the number will be low (for me it was < 1%!).
Sometimes it’s more useful to define what your audience is not, rather than define what it is. I call this an ‘anti-persona’. You can find one by following the exact reverse of the personas above (filter for traits that are less common amongst Netflix users).
In our case, it would be:
Mature Male Managers // 125k
Now that we have a good idea of the personas that use the service (or at least the ones we want to focus on), our next step is to define the problem the service is solving for them. Once we have that definition, we can explore the map of competition for solving that problem (it’s wider than you think!).
// Job to be done: Find a Movie to Watch
On the surface, the functional task is to find a movie to watch. What are the triggers that proceed this need?
Now think about the last time you were in any of those positions… what was going through your mind?
Anxiety:
Inertia:
To drill down further into the ‘job’ And Chill is doing, we can use a simple framework called the ‘five whys’ — it works like this:
1) Q: Why use something like And Chill? A: To find a movie to watch.
2) Q: Why watch a movie? A: To be entertained for a few hours.
3) Q: Why is a movie entertaining? A: It helps you relax.
4) Q: Why is it relaxing? A: It immerses you in an alternative reality.
5) Q: Why does that work? A: You’re escaping from your own reality.
Higher order job: Escapism
So now we’ve arrived at the higher order job — escapism. Armed with this knowledge, we can safely catalogue all the competition (close and tangential) that satisfies this basic human need. If you mapped this out from the first producers of the underlying product, through all the middle men/women to the final customer, what would that look like?:
A couple of things jump out of doing this type of exercise — And Chill’s closest competitors are friends and family. They’re who normally give such personalized recommendations, so AndChill’s friendly human-like bot approach makes absolute sense.
On the other side, less personalized but more widely appealing are the publishers and curators — there is a wealth of content online around what movies to watch. This is a game that And Chill could consider getting into as it’s more personalized data and the fact that it doesn’t need to plaster the content with ads (apart from for it’s own service) may give it an advantage.
The tech titans feature heavily in this ecosystem, from Google search, to the iTunes store and Amazon’s prime video, they’re all investing heavily. This can actually be a good thing as it provides many different routes for acquisition, and puts a large value on the industry as a whole (two things every investor wants to see). Through their deep pockets and treasure trove of data they could of course build a ‘recommendations bot’. However it would be a huge departure from their current strategy and they would most likely opt for buying an existing solution that gets traction.
The content streaming platforms themselves are of course incentivized to solve recommendations. However I suspect they will be too busy betting on developing original content. This solves a problem for them (users will stick with a service regardless of how poor the recommendations, if it is the only place to see their favorite shows) but it doesn’t solve the ‘what to watch’ problem (and in fact magnifies it) for consumers, who now need to trawl through an ever expanding content library across multiple platforms.
In the above graphic we placed a few examples in each area. Now we’ll dive a bit deeper into a full list of the players in each section, including some stats on what channels drive their traffic (from SimilarWeb). This list will come in handy when we start building target advertising audiences.
Streaming Platforms
Publishers
Rating sites
Tech / Apps
Movie Critics
Other Streaming
Cinema Listings
Social Networks
– Facebook: 85% of social traffic based on SimilarWeb sample data
– Reddit: 5.5%
– Twitter: 5%
– YouTube: 2%
– Instagram
Personal Network
Cable Companies
Hardware
Film Studios
Advertising
Retailers
The biggest thing that stands out here is the dominance of search. The vast majority of sites had SEO as their number one marketing channel driving traffic. Our SEO and content strategy will be of utmost importance. Those that didn’t have organic search as their major source of traffic, had most coming from Direct. This is an indication of a strong brand — either built offline or through TV ads (larger production houses / platforms), or through word of mouth (smaller apps / tech products). After compiling this list we might want to dig further into any anomalies, for example how does Film Freak get so much traffic from social (35% — their largest channel)?
To round out this competitive analysis, sometimes it’s useful to take some time to think about what products complement the ‘find a movie’ job, and to catalogue what products did the ‘escapism’ job in history (and what will do it in the future). No need to draw any insights immediately, but these may give you more creative ideas for marketing strategies.
Complements:
Job in history: opera, ballet, theatre, ballads, campfire stories
Job at present: restaurants, comedy, bars, concerts, video games, porn, reading, traveling, sports
Job in future: 3D movies, augmented reality, virtual reality
Now we know who we’re targeting, and who we’re trying to beat, we need to decide how to beat them… with ad-copy. Defining a value proposition is the most important part of a marketing plan — testing proposition variations until you find the right one is the highest value thing you can do to improve marketing performance.
What are the use cases?:
– Date night
– Party
– Solo
What could we call the product?:
– “picks”
– “recommendations”
– “finds”
Time:
Cost:
Quality:
Pain:
Access:
Status:
Proof points:
Note that now we’ve brainstormed these different components, it’s relatively easy for us to put together ad-copy for whatever we need. For example ‘Truly Personalized AI Movie Recommendations that the Webby awards called “worthwhile, indispensable, or even life changing”’. This is a particularly effective document to share with designers and copywriters — it helps them stay on brand and saves them time (saving you money).
We’ve already seen the importance of an SEO strategy through our competitor research. Now it’s time to dig out what keywords are valuable. Let’s start by listing the keywords I would search for, off the top of my head, then run them through Google’s Keyword Planner.
Huh, very few of those keywords are even useful — most of them have very little traffic. I expected a lot more than 170 searches a month from “What’s popular on Netflix” and “what’s good on netflix” is way too expensive at $13.70 per click. Well this proves we shouldn’t just rely on our hunches when planning marketing — we should look at data.
So what are the top ten keywords with volume around Netflix?
The vast majority of this seems to be people lazily Googling to log in to Netflix (heard of a bookmark, people?). However it is interesting to see so much traffic for the keyword “New on Netflix” — this is giving us a hint that we might want to address this in our content marketing strategy.
It might also be useful to run the keyword “Netflix” through Google’s trend tool and see if there’s any seasonal rise (or dip) in search volume.
Interesting — for most of the year we see no real trend, but around the 20th of December to January 16th we see a big spike for the holidays. So we need to take that into account with the timing of our marketing plan.
So let’s actively find some ‘bargain’ keywords. Here I’m looking for keywords that cheap enough but still have a decent volume. Note that cpc will also be a good indicator of how difficult it will be to rank for that keyword on SEO — this isn’t just useful for Google Adwords campaigns.
*Movies to watch on Netflix has 22k visits at $6.06 cpc.
*New on Netflix has 90k visits at $9.89 cpc.
*Netflix Login has 368k visits at $6.91 cpc.
We may be able to use some of these keywords (for example creating an article on how to get a free trial on Netflix), but now let’s find some more relevant terms that might be worth paying a bit more for:
Through the process of finding relevant keywords, you’ll inevitably stumble upon keywords that aren’t relevant. It’s a good idea to add these as ‘negative’ keywords so your ads don’t match to them. It can also make sense not to have these words on your site (if at all possible).
Now let’s look at the traffic for each category of keyword and see what they compare in terms of volume and cost.
Brand terms:
Streaming Platforms:
Aggregators:
Streaming:
Cable:
Publishers:
Competitors:
Critics:
Studios:
Directors:
Actors:
These figures will form the basis of our testing on Paid Search, as well as giving us an idea of what keywords we should go for when publishing content. As you looked through the list, were there any surprises? I couldn’t have predicted the relative search volume of the popular actors and directors, I overestimated concerning the app / tech competitors and I had no idea that Hulu was such a big deal. If you even got a handful of these wrong, then that’s more proof that a data-driven approach is essential.
Now that we have these keywords, we can start to prioritize the different types of content we want to produce, knowing what’s easiest to rank for and what’s the highest opportunity in terms of traffic. If we didn’t have this information we’d waste hours creating content and trying to rank for terms that would ultimately not be worthwhile ranking for.
Keywords are just one piece of the SEO pie — we won’t rank for anything if Google can’t easily crawl and understand our website. Of course in And Chill’s case there are no specific content pages (just the homepage), so this analysis won’t be very in depth, but there’s still some things to fix.
First, I like to just stick the URL into the Google Keyword Planner — this will tell us what keywords Google thinks is relevant for our website, and gives us a quick sense check that we’re on the right track.
Hmmm. Not good. Ok well I guess it’s picking up on the ‘browsing headaches’ vibe and the thin content on the page isn’t doing us much good. This will improve over time as we create content and pad out the keywords on the main page.
Next stop is the Website Grader and Page Speed Insights, both of which combined will tell us the majority of the technical things that are wrong with the site. Note — you could pass this straight to a developer, but I’d take some time to Google these terms so you understand their relative importance, otherwise you might end up paying a lot of money / wasting a lot of time to fix something relatively unimportant.
Now let’s take a look at the meta tags — these tags are hidden to the avg user in the source code, but this is where Google (and Bing) pulls your title and description from when displaying your site in the search results, so they’re very important to SEO. Additionally the ‘open graph’ tags (og) fulfill the same purpose but for Facebook (and other social networks) in the timeline. We’ll rewrite any of these that are sub par based on best practices.
Old Title & Description:
<title>And Chill</title>
<meta name=”description” content=”And Chill gives you movies for Netflix and chill. Stop searching. Start chilling.” />
New Title & Description:
<title>And Chill: </title>
<meta name=”description” content=”Netflix and chill with the top movies to watch based on our personalized recommendations. Don’t Search Netflix forever: text 213–297–3673 and chill.” />
Link to yourself (stop basic duplication):
<link rel=”canonical” href=”http://www.andchill.io/"/>
Social Graph Tags:
<meta property=”og:site_name” content=”And Chill”/>
<meta property=”og:title” content=”Home”/>
<meta property=”og:url” content=”http://www.andchill.io/"/>
<nbsp><meta property=”og:type” content=”website”/>
<meta property=”og:image” content=”#”/>
*note: this wasn’t present, but we should add it.
<meta property=”og:description” content=”I just stopped wasting time browsing Netflix and text @AndChill for a free personalized movie pick. #NetflixAndChill” />
<meta itemprop=”name” content=”Home”/>
<meta itemprop=”url” content=”http://www.andchill.io/"/>
<meta itemprop=”thumbnailUrl” content=”#”/>
<link rel=”image_src” href=”#” />
<meta itemprop=”image” content=”#”/>
<meta name=”twitter:title” content=”Home”/>
<meta name=”twitter:image” content=”#”/>
<meta name=”twitter:url” content=”http://www.andchill.io/"/>
<meta name=”twitter:card” content=”summary”/>
*note: this also wasn’t present, but we should add it.
<meta name=”twitter:description” content=”I just stopped wasting time browsing Netflix and text @AndChill for a free personalized movie pick. #NetflixAndChill” />
I also checked and we’re already verified via Google Search Console:
<meta name=”google-site-verification” content=”XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX” />
One final thing to note when looking at the technical side of things — the main Phone image on the website is not showing with img and alt tags. This means it won’t be indexed for image search or add to keyword depth, so we should definitely change that.
Link Profile
As well as technical on site changes, we should also look off site — what websites are linking to us. In this case very few are, so instead let’s focus on our competitors — who is linking to them? Once we figure this out, we can potentially work out strategies for getting those sites to link to us, and therefore start rising up the rankings. The tool to use here is Open Site Explorer from Moz.
Netflix
Search Visits: 91m Search Visits
Domain Authority: 90/100
Diversity: 11.7k Root Domains, 223k Total Links, 19 links per domain
Velocity: 828 in last 60 Days, 0.37% growth
Link Examples:
A Good Movie to Watch
Search Visits: 368k Search Visits
Domain Authority: 37/100
Diversity: 153 Root Domains, 967 Total Links, 6.3 links per domain
Velocity: 48 in last 60 days, 5% growth
Link Examples:
Netflixable
Search Visits: 528k
Domain Authority: 39/100
Diversity: 13 Root Domains, 3.5k Total Links, 268 links per domain
Velocity: 1 in last 60 days, 0.02% growth
Link Examples:
Search Visits: 1.3m
Domain Authority: 31/100
Diversity: 11 Root Domains, 355 Total Links, 32 links per domain
Velocity: 704 in last 60 days, 98% growth
Link Examples:
Netflix Roulette
Search Visits: 54k
Domain Authority: 45/100
Diversity: 233 Root Domains, 585 Total Links, 2.5 links per domain
Velocity: 40 in last 60 days, 6.8% growth
Link Examples:
Instant Watcher
Search Visits: 161k
Domain Authority: 52/100
Diversity: 470 Root Domains, 92k Total Links, 195 links per domain
Velocity: 398 in last 60 days, 0.43% growth
Link Examples:
IMDB
Search Visits: 357m
Domain Authority: 100/100
Diversity: 34k Root Domains, 4.6m Total Links, 135 links per domain
Velocity: 31k in last 60 days, 0.67% growth
Link Examples:
Looking through data like this for long enough gives us pretty interesting insights. For example the shocking dominance of Netflix Life and it’s 98% link growth — they’re obviously doing something right! It’s also interesting to see the difference between “A Good Movie to Watch” and “Netflixable” — with just 13 unique domains linking (with hundreds of links each) they’ve managed to drive 500k traffic instead of 370k, but Netflixable’s dominance is slipping — only 1 new link in the past 60 days.
Looking at the articles / pages that link to these sites, we should now have plenty of good ideas for how to build links of our own. Here are a few that I thought of whilst doing this:
Contents:
a) Public Relations
b) Editorial Content
c) Email List Building
d) Conversion Optimization
e) Tracking Plan
f) Referral Triggers
g) CRM Automation
We mentioned in the last section that we’d share posts with influencers… well who are they? Where do we find them? What do we say to them? Well using tools like Buzzsumo and Just Reach Out, that’s easy. Here’s what we have after a few hours digging around:
Matt Brian, @m4tt, Engadget
http://www.engadget.com/2016/02/18/netflix-recommendations/
Graham Winfrey, @GrahamWinfrey, Inc
http://www.inc.com/graham-winfrey/metrograph-theater-betting-against-netflix.html
Ben Popper, @benpopper, The Verge
http://www.theverge.com/2016/2/17/11030200/netflix-new-recommendation-system-global-regional
Paul Sawers, @psawers, Venture Beat
http://venturebeat.com/2016/02/17/netflix-says-going-global-has-improved-its-personal-recommendations/
John Patrick Pullen, @jppullen, Time
http://time.com/4147156/netflix-profile-recommendations-trick/
Colin Dixon, @nscreenmedia, N Screen Media
http://www.nscreenmedia.com/netflix-spends-20x-pay-tv-recommendations/
Angela Alcorn, @angelaalcorn, Make Use Of
http://www.makeuseof.com/tag/hide-spoilers-get-random-episodes-netflix/
Travis Hanneman, @thanneman1, TheChive
http://thechive.com/2016/04/07/porn-stars-recommend-the-best-movies-for-netflix-and-chill-8-photos/
Eliza Thompson, @thompsonplaid, Cosmopolitan
http://www.cosmopolitan.com/entertainment/movies/a57845/mean-girls-netflix-recommendations/
http://www.cosmopolitan.com/entertainment/tv/a56824/netflix-shows-you-can-watch-in-a-weekend/
Peggy Truong, @peggy_t, Cosmopolitan
http://www.cosmopolitan.com/entertainment/tv/a57071/recommended-netflix-shows-binge-gilmore-girls/
http://www.cosmopolitan.com/entertainment/tv/a58190/netflix-true-crime-tv-shows-documentaries/
http://www.cosmopolitan.com/entertainment/movies/a56739/90s-movies-netflix-streaming/
Kate Storey, @katelstorey, Cosmopolitan
http://www.cosmopolitan.com/entertainment/tv/news/a55934/netflix-hack-watch-with-long-distance-friends/
I’m noticing a lot of Netflix coverage from Cosmopolitan… let’s search for more links from them all at once:
Search for: “site:www.cosmopolitan.com netflix”
…and you get the following:
We can see here they’re pushing pretty hard on Netflix content — must be a lot of traffic for these stories. We can test some of these angles with our own content, or reach out to these journalists for coverage.
Now, how would we reach out? The (aptly named) Just Reach Out tool has a pretty foolproof template included:
Email: **@engadget.com
Subject: Re: How Netflix’s global growth is making recommendations better
Hey Matt,
I was just reading your article, How Netflix’s global growth is making recommendations better, and thought [insert a controversial point about an article and share your opinion, pique their curiosity a bit, ask them a question].
BTW, I work at a company where we [insert your one sentence pitch]. We were recently featured/talked about in [insert a recent accomplishment]. We have some news coming out [feel free to insert a few words about the news], would love to share more info with you. Let me know if you’d be interested to hear more.
Thanks,
So here’s our pitch for And Chill:
Email: **@engadget.com
Subject: Re: How Netflix’s global growth is making recommendations better
Hey Matt,
I was just reading your article, How Netflix’s global growth is making recommendations better, and thought it missed a key point that I wanted to bring to your attention.
As we saw with House of Cards and The Killing, adapting niche local content for a global audience is a winning strategy. If this new algorithm allows Netflix to connect the dots between fans of that niche content and a globally recurring profile of user, they’ll know exactly what content to pick up and adapt into an original Netflix series — this could be huge!
BTW, I’m the founder of a company that provides personalized Netflix recommendations using AI (www.andchill.io). We recently won a Webby Award where they called us “worthwhile, indispensable, or even life changing” and over 80% of users are happy with their recommendations. I’m willing to make myself available for quotes and I have a number of thoughts on the industry I’d be happy to share with you.
Let me know if you’re interested in hearing more.
Thanks,
Jake
Founder
And Chill
Finally, as well as going to the top and pitching journalists, we should also mix in some lower level posting on Reddit. These posts are an easy way to get in front of niche communities that should love what we do, and will in many cases get us in front of journalists and influencers looking for their next story. Here are some relevant threads (found via Just Reach Out):
We’ve already spent a lot of time looking through popular content, but to round this out we should look at what’s shared the most on social media (we can do this on BuzzSumo). Here are a few of the top articles in the space (they’re not what you’d predict!):
14 Facts About Netflix, Recommended For You
Jake Rossen, Mental Floss
14.3k Facebook, 104 LinkedIn, 18 Twitter, 66 Pinterest, 14.5k Total
@metaphysicalmrt (11.6k), @jacobsmedia (5.5k), @secygen_ficci (2.2k), @nieveelyn (1.6k), @radiorob123 (1.6k)
Netflix: How to Use Profiles For Mood-Based Recommendations
John Patrick Pullen, Time
3.1k Facebook, 55 LinkedIn, 314 Twitter, 17 Pinterest, 4 Google+, 3.5k Total
@time (10.3m), @zasursky (105k), @dimitriwtop (6.5k), @abhinavsnayak (1.1k), @aravindbk (10k), @gopalbalaji (8k)
Netflix to stream fake New Year’s Eve countdowns so you can put the kids to bed early
Alyssa Pereira, SFGate
380k Facebook, 9 LinkedIn, 642 Twitter, 0 Pinterest, 381k Total
@sfgate (268k), @petershankman (169k), @mmbilal (43k), @ajitpaifcc (12k)
6 Netflix Tricks You Aren’t Using (But Should Be)
Supercompressor, Huffington Post
158.3k Facebook, 439 LinkedIn, 266 Twitter, 7.1k Pinterest, 166.2k Total
@the420radioshow (22k), @intenseca (20k), @littlemamajama (14k), @acefilmeditors (12k), @dayveesutton (12k)
Netflix Has Hidden Movie Categories And You Can Use A Secret Code To Access Them
Mustafa Gatollari, Distractify
260k Facebook, 1 LinkedIn, 161 Twitter, 284 Pinterest, 260k Total
@mollyringwald (129k), @bzzagentjono (45k), @archivedigger (24k), @suedesays (24k), @adrienneverones (14k)
After noticing these ‘secret codes’ posts, I dug a bit deeper — it seems to be a pretty popular thing!
Wow — we are really swimming in information now. We’ve looked at popular content from influencers, journalists and major publications in the space, and at this point you’re probably brimming with ideas. Now it’s time to get them down on paper. Here are a few of mine:
Evergreen content:
This is content that we produce once but it remains relevant forever. This type of content is key to the efficiency of your content marketing strategy.
So far we’ve strategized a number of ways to start driving people to the website — how do we make sure they come back? The most tried and tested way is through email capture. Once you have an email address, you can continue the conversation and potentially turn one visit into many. The tool I recommend here is SumoMe, particularly their List Builder, Smart Bar, and Welcome Mat apps.
List Builder
A popup that shows when a user looks likely to leave the site.
STOP endlessly browsing Netflix. Get the perfect movie picks.
[ Enter Your Email ]
Get truly personalized movie recommendations, powered by AI.
[ Enter Your Email ]
We’ll recommend movies you didn’t even know Netflix had.
[ Enter Your Email ]
Smart Bar
A persistent call to action at the top of the web site.
80% of our users are happy with recommendations. TEXT 213–297–3673
Webby Awards — “worthwhile, indispensable, or even life changing”. TEXT 213–297–3673
Note — upon texting in the user is asked for an email address.
Welcome Mat
A full page call to action that pops up (in this case) on the 2nd visit.
IF you liked movie X, you’ll love movie Y.
IF you like Netflix, you’ll love our recommendations.
[ Your Email Address ]
STOP endlessly browsing Netflix.
Get truly personalized movie recommendations, powered by AI.
[ Your Email Address ]
With such a simple website, there’s little to talk about when it comes to conversion optimization. Here are a few notes on the major issues:
There are a pretty standard suite of tools we’d normally place on a website at this stage — namely Google Analytics, Facebook, Hotjar and SumoMe. The code for these goes on every page, and this is what they look like:
<!-- Google Analytics -->
<script>
(function(i,s,o,g,r,a,m){i[‘GoogleAnalyticsObject’]=r;i[r]=i[r]||function(){(i[r].q=i[r].q||[]).push(arguments)},i[r].l=1*new Date();a=s.createElement(o),m=s.getElementsByTagName(o)[0];a.async=1;a.src=g;m.parentNode.insertBefore(a,m)})(window,document,’script’,’https://www.google-analytics.com/analytics.js','ga');
ga(‘create’, ‘UA-XXXXXXXX-Y’, ‘auto’);
ga(‘send’, ‘pageview’);
</script>
<!-- Facebook -->
<script>
!function(f,b,e,v,n,t,s){if(f.fbq)return;n=f.fbq=function(){n.callMethod?n.callMethod.apply(n,arguments):n.queue.push(arguments)};if(!f._fbq)f._fbq=n;n.push=n;n.loaded=!0;n.version=’2.0';n.queue=[];t=b.createElement(e);t.async=!0;t.src=v;s=b.getElementsByTagName(e)[0];s.parentNode.insertBefore(t,s)}(window, document,’script’,’//connect.facebook.net/en_US/fbevents.js’);
fbq(‘init’, ‘XXXXXXXXXXXXXXX’);
fbq(‘track’, “PageView”);
</script>
<!-- SumoMe -->
<script src=”//load.sumome.com/” data-sumo-site-id=”XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX” async=”async”>
</script>
<!-- Hotjar -->
<script>
(function(h,o,t,j,a,r){
h.hj=h.hj||function(){(h.hj.q=h.hj.q||[]).push(arguments)};
h._hjSettings={hjid:148872,hjsv:5};
a=o.getElementsByTagName(‘head’)[0];
r=o.createElement(‘script’);r.async=1;
r.src=t+h._hjSettings.hjid+j+h._hjSettings.hjsv;
a.appendChild(r);
})(window,document,’//static.hotjar.com/c/hotjar-’,’.js?sv=’);
</script>
We would also want to track performance when someone signs up, particularly if we’re going to be actively marketing the site. Typically we’d track this in Facebook, Google Adwords and Google Analytics. The code for which looks like this:
<!-- Facebook Lead -->
<script>
fbq(‘track’, ‘Lead’);
</script>
<!-- Google Lead -->
<script type=”text/javascript”>
/* <![CDATA[ */
var google_conversion_id = XXXXXXXXX;
var google_conversion_language = “en”;
var google_conversion_format = “3”;
var google_conversion_color = “ffffff”;
var google_conversion_label = “XXXXXXXXXXXXX-XXXXX”;
var google_remarketing_only = false;
/* ]]> */
</script>
<script type=”text/javascript” src=”//www.googleadservices.com/pagead/conversion.js">
</script>
<noscript>
<div style=”display:inline;”>
<img height=”1" width=”1" style=”border-style:none;” alt=”” src=”//www.googleadservices.com/pagead/conversion/XXXXXXXXX/?label=XXXXXXXXXXXX-XXXXX&guid=ON&script=0"/>
</div>
</noscript>
<!-- Google Analytics Lead -->
<script>
ga(‘send’, {
hitType: ‘event’,
eventCategory: ‘registration’,
eventAction: ‘start’,
eventLabel: ‘form’
});
</script>
However the main issue with the site right now is that this is not possible — it’s the same number for everyone. Therefore we can’t track which marketing sources led to which customers. For this we would want to utilize some sort of call tracking software. This would give each visitor from a specific marketing channel (or even campaign) a unique phone number so we know when they text, it came from that channel.
Alternatively if we switch it so a user needs to sign up via email (my preference is using an Unbounce page), we can more easily tie back performance. For this we’d need to collect a customer id and store it alongside the email, then fire future events using the measurement protocol.
<!-- Google Analytics CID -->
<script>
ga(function(tracker) {
var clientId = tracker.get(‘clientId’);
$(‘#lp-pom-form-XXX #clientId’)[0].value = clientId;
});
</script>
Using the ID, we can then send further actions for that user, even if they happen offline (i.e. through the course of the text conversation). Here’s an example of sending a ‘complete registration’ event for customer ID 12345:
You can send pretty much anything you want via the measurement protocol — for example here we’re sending the first recommendation they got (which would subsequently show up in Google Analytics for you to analyze):
You can even log their responses in Google Analytics, and the final selection they make after several responses:
Once we have people converting (and we’re tracking it), we should think about the next step — getting happy users to refer us to their friends. Again SumoMe features heavily here — they have a number of out of the box tools that will instantly increase sharing.
We could also prompt users to share from within their text recommendations. For example we could use a ‘click to tweet’ link:
A couple of other ideas to explore:
Now we’ll go through the journey when a user actually signs up for the service. The first handful of interactions is of paramount importance to retaining users on the platform — first impressions are everything. We’ll attempt to rewrite the messaging based on our style guide, making sure to land the value proposition clearly in the first few lines. If we don’t need to change anything I’ll just leave the old message up for you to take a look and see if you would change anything.
OLD Welcome message:
Hi from And Chill. Your movie concierge here. Leo won… WHOA.
To get started, can I have your name and email address? We can do some cool personalization with your Twitter handle, so include that if you’d like as well.
NEW Welcome message:
Aaaaaannnnnndddd Chill. You can stop browsing now — I got your back.
What should I call you?
Great, now what’s your email so I can spam you? (jk I wouldn’t spam you, you get enough spam. Amirite?)
So what’s the occasion? Friends night in, Netflix & Chill or lonely binge session?
OLD Engagement message:
Great! Next, give a movie you like and specific reasons why. Or just tell me what you’re looking for in your first movie. We’ll plug in to the Matrix with the info you provide, and will come back with something great for you!
Here’s a good example of what you can say: I want to watch a film about money and wealthy people, but with an edge. I’m looking for something like Wolf of Wall Street, Match Point, Wall Street, etc.
NEW Engagement message:
So what are you in the mood for?
Here’s an example: something like Wolf of Wall Street because I like to see rich people going nuts.
OLD Qualifier message:
Sure — Anything you want to name about what you like about Blades of Glory?
Also, would you like your picks to be on Netflix only or is outside of Netflix OK as well?
OLD Progress message:
Got it! Searching for you now…
Gotcha — One more sec!
Ah! OK. One more moment!
OLD Recommendation message:
Kingpin is a little more cult comedy-ish, maybe a little more challenging, but could be good for you. (link to netflix) (link to Youtube trailer)
Mr. Deeds may be a more direct pick. (link to Netflix) (link to Youtube trailer)
Do either of those look good?
We’re digging deep into the well to try and make sure we find something you haven’t seen; is Someone Marry Barry something that looks like you’d watch? (link to Netflix) (link to Youtube trailer)
Goon would also be a good pick (link to Netflix) (link to Youtube trailer)
How about Chef — a sweet cooking comedy with Jon Favreau. (link to Netflix) (link to Youtube trailer)
For something light with attitude go for People places things (link to Netflix) (link to Youtube trailer)
NEW Recommendation message:
Mostly great, we should just change a few things:
The following are brand new messages we might want to consider:
NEW: Occasion qualifier
Am I recommending just for you? Or do you have company?
So what’s the occasion? Watching alone or with friends?
NEW: Content qualifier
Just checking — you cool with swearing / nudity or no?
Let me know if you want to avoid NSFW content.
Retention
Now we’ve been through how to greet new customers, what do we do to keep customers from slipping away?
‘Slipping away’
‘Pick of the week’
‘Anticipated need’
‘New arrivals’
‘Y based on X’
Contents:
a) Google AdWords
b) Bing Ads
c) Google Display Network
d) Facebook Ads
e) Twitter Ads
At this point we’ve set up all the basic pieces necessary to grow our organic traffic, convert them into users and keep them retained. Now it’s time to throw gasoline on the fire and run some paid advertising. Given the volume and breadth of relevant search traffic, Google Adwords makes a lot of sense to focus on. However as a new product concept we should test pushing the concept in front of the right people via Facebook Ads.
Ads for “Netflix” keywords:
Ads for “Competitor” keywords:
Ads for “Actor” keywords:
Ads for “Director” keywords:
– Tarantino and Chill?; The one Tarantino movie you haven’t seen? Y.
– Personalized Movie Picks; People who like Clint Eastwood, also love Y.
– AI Movie Recommendations; People who like Steven Spielberg, also love Y.
Ads for “Streaming” keywords:
This is nice and simple — if Google Adwords works, we just upload the same campaigns directly into Bing and we’re live.
We shouldn’t expect this to work as well as the other channels we’re testing for pure prospecting (finding new users), but it can work from a retargeting point of view (bringing users back who visited the site but didn’t sign up). We just need to make a couple of ads (I use Canva) in the following sizes:
300×250
728×90
160×600
If we do want to test prospecting on the Google Display Network, here are a few websites we can buy traffic that we should test:
As a reminder, here are the different audience personas we came up with initially. Because they were built using Facebook’s audience insights tool, we can actually upload them and target them directly. This will give us a good idea of which audience type is most likely to perform.
Millenial Couples // 375k
GenX with Kids // 275k
Blue Collar Single Men // 375k
We should also take advantage of Facebook’s vast likes and interest targeting capability and target people who like our various categories of competitors and complements.
Directors
Actors
Aggregators // 2m
Movies
Similar to our Bing strategy, we should initially just attempt to replicate the Facebook audiences that work the best on Twitter. Although this approach isn’t perfect, and we could get more advanced as we grow, most startups would do better to spend more time getting the main two channels to work rather than branching out too early.
Contents:
a) Time Costs
b) Revenue / ROI
c) The Goal
Now we have our work cut out for us, it’s time to figure out what all this will cost! Check this link for a full budget forecast spreadsheet with cost estimates for all the above outlined activity for the next 6 months.
You’ll notice that a bulk of the costs are related to people. You might say that Jake would do the bulk of this activity himself, and you’d probably be correct for the majority of startups. In fact, a lot of startups I’ve talked to even hire a marketing person but don’t account for their salary when making budgeting decisions.
I think this is a dangerous path — whilst marketing may not always break even for you in a measurable way, and you may not have the cash to front for outside help, your goal should always be to track and assign value to the time you (or your team) is spending.
It’s easy to fall into the trap of working 80 hours a week and hustling in the early stages of a startup and completely forget that your job as a founder is to remove yourself from the equation. Until you can define, standardize and assign value to the tasks you’re doing, you won’t be able to automate or delegate them and your business will have a huge bottleneck… you!
Worse, failing to account for time causes you to choose the wrong marketing strategy in the first place. Failing to account for time biases you against paid advertising tactics that might actually bring a higher ROI than so called ‘free’ organic tactics. If every founder was disciplined in valuing their time would it really be cost effective for them to spend all day writing blog posts, organizing events, posting on twitter and answering Quora questions? It might be a yes in your case, but do yourself a favor and track it to be sure.
People costs in this case are recorded across the board at a rate of $30 an hour. In practice the price would vary by tasks — it could be much more or much less than this, so adjust accordingly in your own models. Remember this doesn’t have to be perfect — simply being smart enough to account for time in the first place puts you in the top 2% of marketing teams I’ve met!
Now this is a tricky one as And Chill doesn’t currently have a visible revenue model, and may not plan to have one for a while… so how do we justify this $50k investment in marketing without a plan for making the money back?
Well a startup is a series of experiments — you take an idea and validate every assumption about how it could be a profitable business, starting with the riskiest assumptions and iterating until you’ve worked out the kinks. Well the biggest risk with any startups is that nobody will care enough to use it. Or more precisely, people won’t care enough to be cost-effective to attract to the service, vs the amount you stand to make from them.
So for example a business with an advertising revenue model may only expect to make $3-$10 per customer per year by showing ads — they aren’t going to be a valuable business if they can’t attract customers for less than that. That’s why most ads supported businesses (Facebook, Google, Yelp) rely on virality and organic ‘free’ traffic to grow. Although both of these have real costs in terms of product development, get it right and the platform can grow exponentially with 1 engineer bringing in millions in revenue.
So my advice when figuring out the revenue component for free apps, is to decide what you eventually want your business model to be, estimate what a customer would be worth in that scenario and then optimize your marketing accordingly until you can acquire customers for less than that number.
The truth is, even if And Chill was making revenue, I wouldn’t expect that $50k to pay back. That’s not a realistic goal for the first few months of marketing for a brand new, unheard of product. The goal here is to learn.
We want to see what makes this business tick — can we get anywhere with SEO? What types of content are getting the most traction? Can we get any press coverage? Are we seeing what value propositions work with paid ads? Have we tested what designs work on the landing page? Have we managed to move the needle on retention?
All of these questions and more can be answered for a relatively small investment of time and money. At the end of the process Jake will know so much about the viability of the business that he’ll either be in a strong position to raise a new round… or he’ll be convinced the idea will fail. A strong marketing plan has the power to make or break the business — anyone who doesn’t go through this exercise is flying blind.
Sign up to our newsletter to stay up to date with all the latest movements in the field.
More from Ladder Blog
The art world is being disrupted by generative AI, and artists aren’t happy. Generative AI models like DALL-E, Midjourney, and Stable Diffusion were trained by scraping millions of images from the internet, without permission. Artists claim this violates copyright, while the companies training these models argue this falls under fair use.
Read More →In the world of digital marketing and data-driven decision-making, creative testing is a pivotal tool in achieving business growth. Gone are the days of relying on gut feelings or guesswork; now, business decisions are powered by data-validated insights, meticulously collected, analyzed, and validated. This transformative process empowers businesses of all sizes, from established enterprises to budding startups, to thrive in an ever-evolving digital market. This article looks at the practical applications, challenges, and innovative transformations associated with creative testing, offering you valuable insights and actionable strategies to implement in your own digital marketing efforts for achieving growth and success.
Read More →
