Module6 Mediabuyingblitz
Module6 Mediabuyingblitz
Module6 Mediabuyingblitz
Traffic BlackBook
Direct to Site Banner Buys - Both Small & Large Sites
Today’s Agenda
- Discuss how to find relevant sites to approach for banner
buys
- One of the ways to approach these sites
- If your CTR are too low, you could end up paying too much
for clicks. Avg. CTR seem to be 0.35-0.55% but this can vary
wildly.
Why This Route and Not Large
Display Networks?
- Larger networks such as Value Click and Tribal Fusion can be
great since you can hit a whole network of sites with your
budget, but your transparency is limited, and testing typically
costs more to be done properly
- With the direct method, you can learn how to do proper
research and media plans
- Again, we are building a proof of concept. Building our own
“network of websites” to get a good profile of what kind of
sites our offers are working on. Once we have that data, going
to a network with a good rep is a lot easier since we know the
kinds of sites we are succeeding with, and can buy targeted
traffic then
Why This Route and Not Large
Display Networks?
- Costs less to build data
- Look for sites getting less than 20k-30k uniques per day
if possible. One of my best small buys had 2500 US
uniques per day.
- You can also search for “related terms” as we did with Facebook interests.
Both methods have given me success.
- Go to Big-Boards.com >>
Step 3: Contacting Sites
- Contacting small sites can be a challenge
- For forums, if none of the above work and you really want to
get on there, post in the forum on how to get a hold of an admin
or the site owner. I’ve had to do this before for a forum in GCN
that was converting great for me and I wanted to cut out
Google.
Step 3: Contacting Sites
Sample Script: This is typically how I e-mail smaller sites...
Chad Hamzeh
DSV2 Media
This is a template I pretty much copy and paste since we are
contacting so many sites. Sometimes I don’t even put their URL in the e-mail. This is a numbers
game.You could even outsource this if the person knows how to select good targets.
Step 4: Deciding What to Pay
Like I said, the majority won’t contact you. If you like, you can e-
mail them several times. If you get a response, a lot of times it
will be vague, and sometimes they will want you to name a price.
For smaller sites, I will typically then ask how many impressions
my ad will see in the month. Tell them you want to buy the entire
ad slot for the month, and not be rotated with other advertisers.
Cross reference their impression number with the total page
views number in Google Ad Planner (if available) assuming you
purchased 100% impressions.
Also, keep in mind that if you are promoting CPA offers, they will
only be eligible for certain countries, take this into account when
measuring traffic levels. This is extremely important.
Step 4: Deciding What to Pay
If we have previously targeted this URL on GCN, or we decide
to do so now, we can get an idea of what to pay to maintain
profitability. For example...
That being said, I am willing to replace the Campaign Cost/page views x 1000 = CPM
Adsense ads on 50% of the page views, with your banner ads, for $400 a
$400/112500 x 1000 = CPM
month.
$3.55 = CPM
Currently I receive 130,000 unique visitors a month and over
225,000 pageviews. I will rotate a couple banners if you wish. I can cloak http://www.clickz.com/cpm_calculator
the links if you want, can link to you with whatever URLs you want, and
will include tracking code if you wish....”
Although $3.55/$400 was too much in this case, depending on what I was
paying on GCN, I might be able to go over my calculated max CPM and
remain profitable, but it’s not advisable. In many cases you won’t have CPM
data from a GCN campaign, so stick to your thresholds.
Step 4: Deciding What to Pay
So in that example I had the luxury of knowing my previous placement data from GCN.
When I don’t have that data, I will calculate a theoretical maximum CPM based on page view
data, a lower CTR of 0.25%, and a conversion rate from the click of 1% when using a lander, and
of course my offer payout. I can use a higher conversion rate for offers not requiring a credit
card. But again this is simply giving us an estimate. Also take in mind direct linking vs. landing
pages.
I will also use CubeStat.com for a rough estimate of the sites revenue if I don’t have a better
option for estimating.
And finally, sometimes I will simply throw out a low figure for these small sites, something like
$50-100 for the month and see the web masters reaction. Worst they can say is no.
As a guideline, I am usually looking to pay what would equate to under $1 CPM, ideally 50 cents
or under. But of course this can vary.
Step 4: Deciding What to Pay
If you’re a bit nervous about one of these buys, they can usually
be negotiated for a term of 1-2 weeks as well to test.
Step 4: Deciding What to Pay
Don’t only take the web masters word on traffic volume and
demographics, do your due diligence and make sure you keep
geographic restrictions in mind.
Step 4: Deciding What to Pay
Try to get a 300 x 250 slot above the fold if available. Followed
by a 336 x 280, 160 x 600, and 120 x 600. Testing is the only way
to know for sure, but I have noticed higher CTR on average from
300 x 250 than any other.
Since we’re paying flat rate, test going direct to the offer and
landing page. Test different factors in this low risk environment.
Step 6: Tracking Small Direct Buys
Tracking can be done in couple ways for small direct buys and I
have done both:
1) Prosper202
2) An Ad Server
Step 6: Tracking Small Direct Buys
When using Prosper202 for small direct buys, each banner you give to the web master will need
it’s own destination URL associated with the correct banner, so you can see which banner is
converting and which are not. Then it is simply a matter of getting the web master to pull the
non-converters. This works well if you are testing and paying for multiple placements, so that you
can focus your funds only on the converting placements.
Step 6: Tracking Small Direct Buys
The other type of inventory is premium inventory. With premium inventory you are usually given
the option to target a specific demographic, genders, ages, or even categories on the site. Premium
inventory is great if you already know your cost per acquisition (http://www.clickz.com/
cpa_calculator) and eCPM, and want to buy from a larger site with broader demographics. This way
you can specify the exact demographics you need.
However, since we are going after sites that already have a specific demographic (75% women
between the ages of 18-24 for example), then we already know that traffic fits our demographic.
And in many cases, remnant traffic will be 2-4x cheaper than buying premium.
At a Glance:
- You will be optimizing the campaign, not your rep
Taking advantage of remnant traffic first is important when testing new traffic sources. Our goal
here is to calculate our eCPM (effective cost per thousand) for that traffic source. In simple terms
this tells us how well that traffic converts for us. Basically, it tells you the most can & should pay per
1000 impressions. Once we have this number after a test run, we can look to renegotiate with
confidence, and many times you can get cheaper traffic by buying inventory in advance. Keep in
mind, we will want to know eCPM per offer, ad, etc., if possible for that traffic source. You will be
looking for the offer and ad with the highest eCPM.
http://www.csgnetwork.com/ecpmcalc.html
At a Glance:
- You will be optimizing the campaign, not your rep
eCPM is kind of an odd metric for advertisers. Typically most definitions and calculators are tailored
to publishers (site owners) It tells them basically how much they are making per thousand
impressions for that ad, etc. The reason we want the ads and offers that bring the highest eCPM, is
because we basically have more room to negotiate since it’s bringing in lots of revenue. Make sense?
If you are hitting an eCPM of $2 on remnant traffic then that is much better than an ad bringing an
eCPM of $0.50. If $0.50 is the max cpm you can pay for that traffic, it is going to be tough unless
you have an extremely cheap traffic source.
http://www.csgnetwork.com/ecpmcalc.html
At a Glance:
- You will be optimizing the campaign, not your rep
The other possible downside is if there is someone buying that sites premium inventory, then it’s
possible you could be left with the demographic you didn’t want. That’s why it’s important to find a
site that has a really dominant demographic for this first strategy.
I will also be explaining a strategy for broader demographic sites, but that takes more experience
than what we’re doing here.
At a Glance:
- You will be optimizing the campaign, not your rep
The other possible downside is if there is someone buying that sites premium inventory, then it’s
possible you could be left with the demographic you didn’t want. That’s why it’s important to find a
site that has a really dominant demographic for this first strategy.
I will also be explaining a strategy for broader demographic sites, but that takes more experience
than what we’re doing here.
Our first strategy will involve choosing 1 vertical and 1 high converting offer.
Again, ask your AM for display data if it’s available. Have 2-3 offers to
rotate if available and budget permits.
I’ll skip past how to select an offer for
this. Same rules apply as before.
- They’ll usually ask a bit about your company, where you’re from, small talk stuff. My buddies say I
lack in this area. ;)
- They’ll ask about your product, your campaign, even your offer brand name. In my experience
with direct sites, this isn’t because they are going to disapprove it because it’s a re-bill or
whatever, it’s because they don’t want your product to compete with an in house product or with
another advertiser sometimes. I’ve never run into an issue here.
- You will want to ask questions. Remember just because you have them on the phone doesn’t
mean you must buy from them. Find out if they have experience with direct response advertisers,
and how other direct response advertisers in your niche have done. Do they have any live case
studies they can show? (this is rare, but a bonus if they will show you)
- Let them know you want to test the smallest minimum they have. Like I said, minimums will
usually be in the 3-5K range. But you get what you negotiate for, and I have been given a 1k
minimum just for asking, from sites that stated they had a 3k minimum. Which leads me to....
- They’ll usually ask you what your monthly budget is going forward. I always say that I have an
unlimited budget assuming the traffic converts to my offer, and that if the initial test looks good
we’ll look at locking in more inventory. This can help your rates going forward. This is especially
important if you deal with networks later on.
On the phone & negotiation
Common points of discussion on the first call:
- Negotiate the lowest possible rate. If you don’t know what that is, it can’t hurt to straight out
ask what the lowest bidder is paying. Obviously they can tell you whatever they want at this point.
- Ignore the prices in their media kit. If you downloaded their media kit before contacting them,
you could see some crazy numbers. I had one buy where the media kit said $27 CPM! I was able
to almost immediately get them down to $9 CPM, which still was too high. (probably why I got
them there so easy). These days I look for a max of $2 CPM, and preferably lower.
- One thing a few guys have said is that there are 2 rates, one for brand advertisers and one for
direct response advertisers. Brand advertisers are usually massive companies not really looking to
generate profit necessarily, but looking to get their brand and name out. This can be effective if
you have your own product of course, but requires a budget. Let them know you are a direct
response advertising agency and you’ll be tracking ROI closely, and that you need a minimum rate
test to establish your eCPM, and if that’s successful, you’ll consider buying much more inventory
afterwards, maybe even premium inventory if it makes sense. As mentioned above with the media
kits, I’ve found this “2 rate” phenomenon to be true.
Insertion Orders
The insertion order is basically a document which specifies the details of the advertising campaign.
Below is a sample IO. Usually provided by the publisher but you can also provide your own.
- Out clause. Your out clause is the amount of time the publisher has to cancel the impressions and
the IO. Really work to negotiate a 24 hour out clause. Talk like you expect it. This way even if you sign
a large IO, you can get the traffic stopped within 24 hours of your request. Never go longer than 72
hours here. I’ve only done 72 hours once, otherwise it is 24 hours. When exercised, tell them to pause
the traffic immediately. Even if your out clause is longer, often times they will pause immediately if you
ask and have specified this in the IO.
- Bill of your ad server. This is very important. In most cases with large publishers you will need
an IAB compliant ad server to be allowed to bill off your own numbers. I’ll talk about ad servers more
later, but it’s important to bill off your own numbers. Many won’t allow more than a 10% variance, but
I haven’t noticed much more variance than that yet.
- Even delivery. Again, very important. This ensures that impressions are delivered evenly over the
term of your buy (30 days, 60 days, etc.) Without this, they are fully capable of dumping the remainder
of your impressions if you exercise an out clause. Another point about even delivery. If your term is
for 30 days, and for whatever reason you need to pause the campaign, let’s say for a week or so, when
you return to the campaign, your 30 days doesn’t get extended by a week. So, what will happen is since
their server needs to deliver the set amount of impressions within 30 days, it will try to make up for
the week the campaign was paused. This happened to me, I was spending on average $100 per day and
had to get it paused for a week, when we returned we were spending much more!
Insertion Orders
We want to test small for the most part. These are the important terms to have in your IO:
- Pause Immediately. Many people don’t know that you can pause traffic in a media buy. Like
anything else it needs to be negotiated and placed in the IO. This is a good one to have in case you
need to pause traffic for an urgent reason.
- Daily Cap. In my experience, they usually can’t hit this exactly but often you’ll get very close. This
is good to have in place especially for the first few days to see how things are coming along. Usually
termed as an daily impression cap.
- Extended Flight Dates. First off, if possible, specify the exact dates of the buy. Now, extended
flight dates isn’t really a term, it’s just another point you can negotiate on. Some sites will tell you their
minimum is $5k per month. However, many publishers will only specify a $5K minimum. It doesn’t hurt
to ask that the minimum be spread over an extended flight date of 60 or even 90 days. I’ve noticed
this to be more beneficial when I have credit terms and I’m not carrying a debt on my credit card.
Structured right, you could be buying traffic from a major site for only $100 per day. That said, at
$50-100/day it will take you longer to get your data, but yes I understand it “seems” safer for
beginners. Start your campaigns early in the week in case something goes wrong!
- Refund unused impressions. If you aren’t getting credit terms and you are paying up front,
make sure it is specified in the IO that all unused spend on impressions is refunded. This is pretty
much an assumption for most reputable sites but it’s good to specify this. Keep in mind refunds can
take a while to get, so get credit terms if possible!
Insertion Orders
We want to test small for the most part. These are the important terms to have in your IO:
- Remnant Traffic. If we’re starting with remnant traffic, specify this and ensure that the agreed
upon CPM is there as well.
- Frequency Cap. With these larger sites I like to test a 1/24 frequency cap, which will limit your
volume at first but we’re not too worried about volume right away. Specify this in your IO. If your test
is successful you can increase the cap which should increase your traffic quite a bit.
- Specify Country. This is so important.You don’t want to end up finding out later that the
majority of your impressions were for countries not eligible for your offer. Sometimes you will need
to speak to a different rep for a different country. If you have your own product, or Clickbank, it’s up
to you to decide where to market. Keep in mind that you should be able to negotiate a cheaper price
for traffic outside of the US.
Insertion Orders
Media Buys really aren’t that risky once you have the right
insertion order in place. Remember, everything is negotiable
and if you have something else you want to add, it can’t hurt
to ask.
Insertion Orders
Getting $5000 On
Credit VS. Paying $3000
Upfront
Credit Terms
VS.
X
Getting $3000 On Paying $3000
Credit Upfront
Credit Terms
- When buying traffic in general, if you can get long credit terms
and short payment terms from your networks or payment
processors, it puts you in a great position to avoid a spending
bottle neck
- With PPC engines, your options are limited. I’ve been in situations
where I was spending $1000+/day on Adwords and didn’t get paid
for 2 weeks, so I had to split the balances between 2 credit cards .
Once I started receiving payments twice a week, this was no
longer an issue.
- With large banner buys, you have a couple options. If you can get
credit terms this is best, but you need to know how to control
your spending and have your IO terms in place. Credit terms are
typically 30 days, but you can even negotiate 60 days!
Credit Terms - Approval
You’re usually going to be asked the following to get credit terms:
- They will also check into the D&B (Dun & Bradstreet) database
Credit Terms - Considerations
A few things to keep in mind in regards to credit terms:
X
Awesome! I got 10k in my account, that means
I can hit up 5 different networks and publishers
to each grant me 5k credit, which MEANS I could
be working with 25k on credit to test! I RULE!
X
Awesome! I got 10k in my account, that means
I can hit up 5 different networks and publishers
to each grant me 5k credit, which MEANS I could
be working with 25k on credit to test! I RULE!
Once the campaign terms are actually agreed upon, it’s time to start the
campaign. For large buys you will be using an ad server. It will be your new
best friend. To actually get the campaign started, you’ll send over ad server
“tags” to your rep. These are typically iframe tags that the site will place in
ad spots which will allow you to externally control ad rotating while getting
stats in real time.
Whew! Back to the Campaign...
I like to make 3 banners per banner size. I make them very different so I can
see which direction to go in. Again, I will want to send 100 clicks or more
to each ad/offer/landing page before deciding what to keep and what to
modify all based on what’s getting the highest eCPM. If an element isn’t
converting I will resort to my desired CPA numbers to decide.
Whew! Back to the Campaign...
Optimizing is pretty simple at this point. The Ad server uses pixel based
tracking so you’ll be able to see in black and white which ads are giving you
the highest eCPM, which are getting the highest CTR, which offers are
converting best, etc.You will optimize just like the sculptor metaphor in our
GCN campaigns. Getting rid of what isn’t working and keeping what is.
Another Strategy
As I mentioned, our 1st strategy is pretty similar to what we’ve done
before. Find a suggested high performer, target those demographics, test it
out. This next strategy is for broader demographic sites, some social sites,
and works great for direct linking. This is one of my favourite but can take
more money to get enough data to make decisions.
Another Strategy
Here we are basically finding large traffic sources and testing different ways
to monetize their traffic. This is similar to how some do placement
targeting campaigns in GCN, and PPV campaigns.
This strategy is basically
about throwing a lot of
mud at a wall and seeing
what sticks. It can cost
more since we are going
to be testing more
variables, but when it
works it can be scaled
quite well.
Mud Slinging: Overview
- Instead of testing 1 vertical with a couple offers, we are now going to test
several verticals at once with multiple offers.
- You will be testing several banners at one time. I like 3 banners per size per
offer. If we have 6 offers, that becomes 18 banners just for one size. See how this
could take more budget to get enough data?
- We want to do this with offers that can be direct linked. Since we have so many
offers/ads to test, we don’t want to add another element (landing page) to get
data for since it will increase our budget more so.You would need a high
conversion rate to get a good enough ROI.
- We want publishers that have the capability to deliver lots of volume, more than
300,000 impressions per day and preferably 750,000+ per day.Verify the number
with your rep based on your CPM, and not just trusting 3rd party sites like
Google Ad Planner.
- I liked to get a CPM of $1 and under. Price isn’t the be all end all factor, but I
find getting traffic under $1 CPM is a bit more forgiving for the types of offers I’ll
be using for this.
Mud Slinging: Overview
Wide demographic offers for wide demographic sites:
- Credit score offers
- Ringtone offers (believe it or not)
- Quiz offers
- Education and continuing education
- Debt offers of all kinds
- Insurance based offers
- Entertainment such as Netflix, Blockbuster, & Pushplay
- Make money and weight loss that have strong pre-sell
landers of their own
- Loan and loan modification
**We will again want to consult our affiliate managers and get the highest
EPC offers in 4-5 of these verticals.You are free to test more verticals/
offers but that means more budget **
Mud Slinging: Step by Step
Step 3: Create all your banners. Feel free to use some of the
network creatives as well since you need to make so many. Load up
your ad server with your offers and creatives, place your ad server
pixel on all your offers, and send off the ad server tags to your rep.
Mud Slinging: Step by Step
- I’ll make the ad server show my highest eCPM profitable offer and
ad 80%-85% of the time, so 15%-20% of the impressions will be
used to test other offers and ads. The 2nd and 3rd highest eCPM
offers will be kept in mind if offer 1 gets pulled or gets stale.
Mud Slinging