There are many changes afoot Pressyo, my startup. This afternoon we were discussing business models, in particular, a deep discussion of the ad-supported business model. It is by coincidence that I work in online advertising as well, so here I will share some insights to the ad-supported business models. I will discuss other business models in future posts* If you’re a long time reader of this blog (one of the two who are not my parents), you may note a change in writing styles. I am writing this on my own free will. I swear Larry Page is not standing behind me with a gun pointed at my head.
Say you want to start a startup. Your investor asks you: what’s your business model? And you answer: oh it’s simple – advertising. You see your investor’s face go from 🙂 to :(. You wonder what’s up.
Your Users Are Products
Here’s what’s up. When you start a project with advertising as its main business model, you need to be very aware that the product is your audience, not your product or service you are building. Your product can be a content based site (blogs for example), a community based site (forums, social networking sites) or a service (a music streaming service for example), but it must be absolutely clear, and there must be absolutely no doubt about it: if you use advertising as your business model, your product is your audience.
Many startup founders I meet often tell me that they want advertising to be their main business models. Very quickly I would be able to determine if that could work. I have asked founders if they are willing to treat their users as products to be sold, and a number would falter. If you’re the kind who prides yourself in the feature of the service/startup you’re creating and you’re unable to reconcile the fact that your users are the product, don’t use advertising as your monetization plan. You’d have a bad time.
OK, so you’re ready to accept that your users are the product. So how would you get started? There are mainly three sides to this business model (and the god help you if you’re in one of those situations where there are two sides to the market): the Publisher, the Advertiser, and the product.
You’re the publisher. Your job is to build a product that meets the Advertiser’s demands. You won’t necessarily have one advertiser or advertiser network. In fact it is more than likely you will be rumning a few at the same time to see which advertiser network can pay you the best rates.
Some Math Required
Speaking of rates, here are some terminologies you need to quickly get up to speed:
- CPM: cost per 1000 impressions. This is the most common method of selling ad space. This means you’re selling a fixed space to an advertiser for a fixed rate of X dollars per 1000 ad impressions.
- CPC: cost per click. This means you’re selling a space to the advertiser, but the advertiser will only pay you if your user clicks on the ad.
- CPA: cost per action/acquisition. This means you’re selling a space to the advertiser, but will only get paid if your user clicks on the ad, and completes an action on the advertiser’s site.
- eCPM: effective CPM. Because ad space can be sold in a variety of methods (three of the most common are listed above) at any given time, the eCPM is a good common-ground metric to talk to advertisers about
To keep things simple, we’ll assume that ads are sold purely on a CPM basis – say you have managed to get a decent deal $0.50 CPM for your site at the beginning. Your revenue is simply the amount of impressions multiplied by the CPM. If you put one ad on a page, it’s safe to assume that it’s one impression that gets loaded when the page loads. If you put three ads on one page, that’s three impressions per page load.
Say you run a simple service – say you run 1 Dyno and 1 Worker on Heroku per month for your service * har har, who does that. You’d typically be looking at a minimum of 2 Dynos and a few Workers. Because you know, your startup is ‘cool’ and you need to be highly available. /sarcasm. . That costs $35ish. To break even you would require 70,000 ad impressions per month. Which is not too difficult, if you have three ads per page – you only need about 25,000 page views per month, which is not difficult to achieve with organic traffic. Given the correct social media and SEO strategies, 25,000 page views per month would take about 3-4 months to scale up. Not a problem.
The Underlying Assumption Is Wrong
However, the underlying assumption is wrong, in two parts: 1) $0.50 CPM is a big ask, 2) you’d have to be really lucky, really well connected, or really charming to have a straight CPM deal when you start off.
You probably will not get a guaranteed CPM deal when you start off with an advertiser network. Instead, you would be given a revenue share. In a revenue share, the advertiser network buys a sells your ad space to a bunch of advertisers via different means – CPM, CPC, CPA etc. The advertiser network gets paid by their advertisers they aggregate, and then shares their revenue with you. This is why it is important to understand the eCPM metric.
Getting a $0.50 eCPM is generally a big ask if you’re just starting out – your product (i.e. your users) would have to be a very unique bunch, and would typically have higher engagement rates than normal – this is due to the rise of automated media buying and better targeting options – I will talk about it in a future blog post.
Naturally, there will be ad networks who will tout that they offer high CPM deals. While I do not doubt they have such deals, I doubt that a startup would be able to secure such deals.
One of the main challenges is getting to the minimum required page views. Take one of the larger * I personally don’t really like them. No reasons why ad networks: Tribal Fusion. Tribal Fusion has a requirement that a publisher’s site has a minimum of 500,000 unique page views a month in order to join their network.
To grow your site organically to that amount of unique page views will most certainly take a lot of time and effort. You may be tempted to buy ads to increase the traffic to your page, and then your business model will have pivoted to the advertising arbitrage model, which I will discuss in future blog posts.
Another challenge that you will most certainly face is audience diffusion. Ads by design are meant to distract the user’s attention to it, and preferably click on and leave the current page. If the user does not spend enough time on your site, then how are they expected to return? Your site becomes a spot for transient visitors. Again, there may be temptation to pivot the business model to an ad/attention arbitrage model.
In The Social Network, Mark Zuckerberg didn’t really want to sell ads in the beginning because Facebook was ‘cool’* In real life, Facebook DID sell ads in the beginning . That is the other factor to take into account. Advertising in your startup will bring in revenue, but will make your site lose it’s cool factor (if there were one). Balancing the cool factor and advertising will be a difficult task.
In fact the whole act of managing advertising is hard. Depending on size and breadth of your advertising space and ad networks you engage in, you may need to run your own ad server. You may not need to either.
Given the above conditions, it is not surprising you find many websites that plaster ads all over the place. It is indeed tempting to think: more ads = more imps = more revenue. While this is still roughly true, by and large, the online advertising industry is moving more and more towards intelligent audience buying.
Real time bidding engines are getting more and more intelligent, and would be able to guess with quite good results, how valuable a user is to the advertiser. This allows the advertiser to only pay high amounts for audiences that matter to their ROI, while paying low amounts to audiences that don’t.
What happens is then this: a segregation of high paying advertisers and low paying advertisers form. When I say high paying advertisers I mean advertisers who would pay high for the right audience. They will pay significantly higher than the low paying advertisers, who would “spray and pray” their ads in hopes of reaching the right audience.
Here, the path of the startup founder diverges: does she want to go the high paying, quality audience route (requiring fewer ad placements), or does she want to go the low paying, low quality audience route, necessitating plastering of ads?
In my personal opinion, the latter is not sustainable business (in fact the only way to sustain it, at least from a mathematical/financial point of view is to pivot into the aforementioned attention/ad arbitrage model).
You Run A Startup, Right?
Let’s say you want advertisers to pay you high rates. This means you have to provide a quality product to begin with. What is a high quality audience? To put it in simplest terms – high quality audiences means that the audience is valuable to the advertiser – either because they are highly interested in the product the advertiser is advertising, or they engage in the ad.
Either way, the way to build a high quality audience is… through your startup product. That indeed, is the function and purpose of your startup product or service: to build better quality audience. I would even go so far to say this: build just enough of your product/service so that you maximize your ad revenue. If you could add a feature that increases ad revenue, do it. Else, you should not even bother building it.
The Optimization Trap
Balancing audience building vs optimizing for ads is a tough thing to do. Many startups fall into the optimization trap. Sooner or later, they will discover that audiences are the product, and they slowly abandon the actual product/service they’re building in order to earn more money from advertising. You see it all the time: shrewd ad placements meant to generate clicks, disguised ads, etc.
More time will be spent optimizing the ad revenue, and less time is spent on audience building (i.e. you know, actually working on the product/service they’re offering). It’s a vicious cycle, really. As you optimize for ads, your audience quality drops, and your payout rates drop. So in response you optimize for ads even more. The cycle continues. Eventually you will experiment with buying traffic in hopes of them clicking on your ads. Welcome to the attention arbitrage model.
There is a lot said in recent times about bullshit metrics. If you run a startup with a product or service with advertising as your business model, what is conventionally known as a bullshit metric, i.e. your pageviews and your unique visitor becomes a proper metric to be aware of.
Instead, I would argue, if you run a business with advertising as your monetization plan, other metrics (like user actions) become the bullshit metrics, until such a time when you can correlate those actions with revenue. Remember, your job is to build the product and sell them to the advertiser, and your product is the audience.
Lastly, I would like to add that advertising is a long game. If you use advertising as a business model and expect instant revenue growth, I would advise you to change your business model and go for a subscription or sale type model.
Advertising as a business model can be profitable if properly managed. Startups that use advertising as a model play the long game – building out audiences, whether consciously or subconsciously over time, before plunging into monetization. Startups that start straight off the bat with advertising are typically not sustainable as they would fall into the optimization trap.
Next business model that I’ll be talking about: affiliates!