How Header Bidding Can Actually Increase Page Load Speeds for Publishers — Talk with Gabe DeWitt

This week on the podcast we talked publisher side latency and header bidding with Gabriel DeWitt, who is the VP of Technical Operations at Index Exchange. Gabe was highly recommended to us as the person to break down the complexities of latency and what most publishers can do to speed their site up.

Like this Podcast??? Subscribe Here

This was an intensely informative podcast, we covered a lot of information on ad tech and how it relates to load times.

The most frequently talked about concept within latency is the affect that header bidding is having on publisher site. Gabe has an interesting perspective on this and doesn’t think that header bidding has to slow a site down. In many instances it can actually make ads load faster.

His logic is that reducing the amount of pass backs that you’re site is running, if done correctly, will decrease overall ad load times, which makes a ton of sense if implemented correctly.

Gabe also made a great point on properly measuring latency. Perhaps the most important is deciding on a measurement technique that you’ll use to determine if you’re changes are making a difference.

Index exchange is always releasing new and innovative products, which you can stay up to date on here: http://www.indexexchange.com/

They also write a pretty great blog, which is rare in the industry, which you can find here: https://kx.indexexchange.com/

Show Notes

2:00 — Intros
2:30 — Gabe’s Background
3:05 — Does header bidding make latency worse or does it depend on how it’s deployed?
3:30 — What are the components of latency? What makes a site slow?
4:30 — How header bidding can increase page load speed
5:15 — What are the low hanging fruit for publishers to increase load times?
5:30 — The importance of using asynchronous JavaScript for header bidding
6:11 — Managing bid requests and bid responses in header bidding
6:43 — Being sensitive to connection speed of your users
7:20 — Filtering amount of bidders by client device speed and type
8:15 — Using time landscape to determine network speed
9:27 — What is the right amount of time to leave the bid window open for publishers?
10:10 — The KPI for Header Bidding should be on total revenue not time or CPM
11:55 — Why slower loading ads can be better
12:40 — The most important thing that you need to measure latency is good and consistent reporting
13:07 — Measuring latency with DFP and Google Analytics
14:15 — Decoupling ad loading from content loading
15:00 — Looking at your loss rate in DFP is a key variable
15:30 — Is it possible to isolate the latency component in revenue vs the other factors?
16:00 — Why eCPM will go down with more bids
16:33 — What stack of tools did you use to manage latency?
16:45 — Measuring key values in DFP for latency
18:27 — The types of assumptions that you can m
18:47 — Rank ordering Header Bidding Partners by revenue, or fill or relationship and call the major ones only for slower devices
19:00 — Set a global time out
19:55 — Are there types of ad tech that aren’t worth the latency trade off?
21:00 — Understanding your audience and their behavior/time on page allow you to ball part their latency tolerance and see what you can work with
21:50 — Publishers should consider header bidding and it doesn’t have to make your page slower
23:39 — Header bidding hits RAM/Network Bandwidth vs waterfalls which hit DOM rendering and is generally slower
24:00 — Where are most publisher now with managing latency?
25:02 — Will header bidding wrappers always be Javascript based?
26:10 — What features still need to be added to header bidding?
One clap, two clap, three clap, forty?

By clapping more or less, you can signal to us which stories really stand out.