As the owner of monetization for your companies web property – you have successfully implemented a multi-partner, programmatic Real-Time-Bidding (RTB) waterfall that’s generating impressive yield numbers. Because of your excellent track record, Management is now expecting you to consistently increase your key performance indicators (KPI’s) – traffic and Revenues ($eRPM).
You know the next couple of months are historically lower (seasonality) and you have proactively communicated this risk guidance to your management and had modeled the decline in revenues with this years financial model. That being said – you know the Company could really use the cashflows.
I am going to take a near and long term approach to Exclusive deals to exploit the traps and pitfalls and how to guard against them. I will also talk about the long term impacts and how to reverse the impacts of a prolonged, negative Exclusive deal.
Then you remember the monthly conversation with one of your networks – who has been pestering you about a rate guarantee with exclusivity. If you take the option, you can lock in your expected revenues and – voila! – problem solved. right?
As you probably know, not all networks are created the same. In my experience – ones that manage brand-safe experiences (like Casale/IndexExchange, WPP Xaxis, ) are less likely to offer wild and juicy deals. Others like Google’s AdX are so dominant that they won’t even “go there.”
Ok, Everything sounds great! What could be the problem? Well, if your traffic numbers stay constant (or grow) and there are no unfortunate changes in overall Ad spend levels (e.g. Post-Tax or Post-Christmas) – then things should be fine. Well, If we thought there was no jeopardy in traffic levels or Ad Revenues we would stick with our finely crafted waterfall. But, we’re not sure – are we? We check the analytics platform and know that historically – traffic drops 20% at this time of the year. So, we call the Ad Network up – tout the unprecedented growth (conveniently omitting the historical trend insights) and get them to commit to an exclusive, rate guarantee for the month.
The first step is understanding the motivations/needs behind the core actors/stakeholders in the arrangement.
- Gets to lock-in revenues they feel is at risk
- Ad Network Sales Manager
- Well, (s)he’s being paid on traffic booked. So this is a short-sighted gold-mine
- Ad Network
- Displaces inventory from their competition
- Additional inventory = additional DSP/Agency buys
Implementing, monitoring and quantifying the lift, lost-opportunity or changes to engagement:
Yes, The best way to approach a guaranteed rate deal is to institute a 90% (exclusive)/10% (existing) A/B test – provided you have enough customer traffic to keep the existing waterfall’s advertisers satisfied. Its important to note that you need to make sure your control group (the 10%) is static so you can accurately measure any potential changes in engagement fatigue. You will want to keep a close eye on the following points:
- Pages/Session (aka Pages/Visit) – This is my favorite metric as it tells you if the users are sticking around and consuming additional content (and generating extra pageviews)
- Time on Site – Again, are there changes in the duration of engagement over the control?
- Bounce Rate – Is the new Advertising causing users to flee the sight? Often advertising that includes auto-play ads will cause this metric to skyrocket.
- Number of Sessions (Visits) per User – Insight into impacts of repeat users
- Traffic levels from attribution sources
Unfortunately, since you cannot easily carry this A/B testing to the Attribution layer (e.g. Facebook, Google, LinkedIn, T.co) – you will have to employ more complicated analysis (like MANOVA or MANCOVA) to try and measure the inherited impacts of UX changes into Referral sources.
Lastly, keep a keen eye on User Feedback (UGC) via site feedback, your call center (if applicable) and Social outlets. In my experience – your loyal customers will communicate here first before leaving. If you see a lot of negative feedback here and it doesn’t wane off – you should expect to see changes in your key KPI’s shortly.
The Cautionary tale:
I have personally seen and experience several situations where a drop in traffic has threatened Exclusive, rate guaranteed deals. It is under these circumstances that the Ad Network/Exchange will call up you as the business stakeholder to threaten the pull the deal unless you make certain concessions to increase the monetization effectiveness of the property. This can range from removing/altering block lists, changing ad unit placements, adding additional units, and lifting bans on more ‘contentious’ technologies. These all (in varying degrees) have a negative impact on the UX and they will hurt your overall traffic in the long term. The following is a typical scenario of how this happens:
Q0: Equilibrium. This is the starting point of the relationship
Q1: The Ad Network notes that your supply of inventory (impressions) has diminished due to the seasonality. This jeopardizes their profitability with the guarantee. In order to maintain equilibrium of Price level ($), You need to either increase your supply (adding more units) or by increasing demand (by adding additional ad technologies or lower quality inventory). Otherwise, the Network will have to cancel your deal. You sell the issue to your management team and they remind you you've already committed to accruing those revenues.
Q2: You reluctantly agree and the Network adds the previously restricted advertisers and/or technologies. The net price level ($) is back at equilibrium (Q0) levels, but this has introduced an untenable UX which will be impacted in Q3.
Q3: Within days you notice a spike in customers complaining via social media, call center feedback, email comments, et al. User engagement metrics start to wane. Site load times begin to grow, especially on mobile. Over the next month(s) you notice more leftward shifts in Demand without the ability to counter with Supply adjustments. Here are the tell-tale symptoms you will experience..
- Bounce rates have climbed up
- Page/Site load times are now noticeably longer than originally and your site loads slower than your competition
- Attribution from Google, Facebook and other prominent sources had become unreliable and wanes heavily
- These sites value the user experience as measured by clicks and bounce rate. Bad site UX causes clicks to drop and bounces to go up. Since this negatively reflects on Google and Facebook, they will organically showcase other properties instead.
- Traffic has fallen outside of your annual model’s risk variance now
- t this point, the traffic and engagement are below the networks tolerance levels and they will either
- Ask to add additional Supply (Ad units) to the experience to maintain the rate guarantee against declining traffic
- Reduce the rate guarantee
I have seen the destructive impacts of this cyclical behavior ruin websites. Once you get to point #6, you have a untenable situation and you have broken your relationship with the most important stakeholder, the customer. This is why it is so important to conduct A/B (90/10) testing in parallel with an Exclusive deal. It not only lets you know when traffic is being negatively impacted against the control, it also lets you know when the control is out-performing the guarantee (so you can call them up and ask for a better rate). The worst possible thing you can do it sit idly by and let the exclusive network dictate your business.
Reversing potential long-term damage:
Over time, UX degradation will result in a decline in engagement, visit frequency, and attribution. Since these are the driving factors for Pageviews, and subsequently eRPM – we need to regain the traffic (and the trust of the users). This step is painfully simple. In order to reverse the negative engagement trend – you will need to level-set the user experience to pre-exclusivity. This includes removing additional Ad units, removing the tainted ad network, and returning to a waterfall of brand-safe advertisers. Most importantly: you will need to apologize to your customers and invite them back. Unfortunately, the cold fact is – it takes longer (and with greater expense) to travel up-market than it does to travel down-market.