本帖最后由 plsd2012 于 2016-10-8 16:22 编辑 p# o% c' n$ {* g8 u7 Y: c- `8 y6 Z3 ]5 \$ A' x# H: x* {2 n
好久没来论坛了，今天分享一篇offer的流程，是业内大牛Malan在他的私人forum分享的，看过以后感觉他的测试方法很系统化，值得借鉴。我就懒得翻译了，悄悄搬运过来分享给大家，希望对新手朋友有所帮助，老手就可以略过了。 y. \2 O. W' i; B( v* h" [* [/ b ; m+ @! Z; I+ A$ `3 M* \3 J# H一个好的offer是盈利的基础，所以测试offer是aff大部分时间在做的工作。如果能有自己的一套测试offer的流程，不但能把损失降到最低，还能尽快测出好offer。可以参考Malan的测试方法，并结合自己的budget情况，适当调整。他是拿Native流量源举例的，任何流量源，任何联盟，任何vertical，都可以参考。4 M! ]# Y7 q" w; X1 A. _
8 r6 t; q& i b# t
----------------------------------" g: P: ~6 ^6 V) [$ `8 W
图片挂了，懒得传了，不影响理解上下文。 " A6 W9 g+ @& g' j ( x2 {% O' P$ g- E8 L
The Beginning / 3 Campaigns) n3 Y) \7 H. T, G
6 S, x6 C! `, i7 T* x5 f0 l! uWhen I test new campaigns I like to start by testing 3 campaigns at the same time. These will be 3 campaigns that I find already running on the traffic source.3 i4 g* U) d1 b- ]/ [
& L1 L/ T8 [" j0 MFor me, this is much faster and more efficient than going through the entire process 3 times to find a winner. 5 Z. M6 R+ a* s6 k \- k0 h * V1 K' p! ^+ |$ ?( l4 l. VImagine spending 3 weeks testing Campaign #1 only to find it doesn't work, then spending 3 more weeks on Campaign #2 to find that it doesn't work and then finally getting to Campaign #3, which works... but it took you 6 weeks to get to it.+ ?; T" O% _/ |& W, e
9 G- }9 J; t3 K3 w2 V
Testing 3 campaigns at once saves you all that wasted time. So take 3 campaigns that are already live on the traffic source and we'll quickly see which campaign gets the best reaction.; J$ k) V, w/ V! s
& \ f" w4 K- o9 j/ c5 \- ~' b
In this example we're going to test Auto Insurance, Skincare Product and a Golf Swing VSL. % d$ G! w; `8 z- \1 J) G) D3 B9 G7 e
Each campaign will be made up of 3 ads. These can be made up of:* s. [# v8 I* t8 M
Two ads being used by others, and one of your own.
One image being used by others and two of your own.
Three being used by others.
Three of your own., ~$ V. D h- n2 ]& D" s
I prefer to mix my ideas with existing ideas. (Either option #1 or #2) : T j+ X; p' b. [' \! f7 T8 S5 G0 h8 x4 ~8 \( C- Y6 s Here is what the 3 ad groups look like. (This example uses one headline/body copy but you could definitely test different headlines/body copy in each ad if you wanted to.) 5 V, J8 a* G0 J' o- v9 X5 W , {* y! d: Q5 e1 O! V[img]http://madsociety.net/forum/attachments/step1-gif.197/对不起，我们不允许远程图片. 请将图片上传到论坛 7 t! L( V- d; m$ ]4 ] f: a
) |0 \+ z3 H1 b9 G5 T1 DEach campaign will use one landing page for now. Preferably one that is proven and being used by others on the traffic source already.; d7 L' p" O! v6 a4 w/ @0 C5 {7 |
! y0 W; [, x" k. |2 J" m
(If it's a direct link campaign, you will obviously not be using a landing page) 8 Z" r# r2 Q* D8 \3 H5 P1 a- l4 ` 5 f7 j+ P3 \- P! ?9 M8 [2 TEach campaign gets a budget of 1x it's offer payout. In this example we have: m, ]+ X$ F3 O% X
Auto: $6 payout - $6/day budget
Golf: $25 payout - $25/day budget
Skincare: $40 payout - $40/day budget : @% n e* x0 A
Run these 3 campaigns for 3 days. $ b6 H* h5 z. V3 Z: d9 _' d# A ; w( v# d& I* SAfter 3 days, choose the best campaign based on performance./ Z! n9 e7 l1 G. q7 o! R" }
6 L- k0 {5 I! i. l, n5 Q There are 3 main performance indicators/ b: b/ n; i# `, A
Ad CTRs
Landing page CTRs
Conversions ' c* T! B7 L( w( V G/ T# v
Whichever campaign does best on those metrics is the one you'll want to go for.; Q# B2 u8 `8 r1 z( |
3 r' N1 J7 ^& S4 U- M9 ]& JExample Results after 3 days:7 \+ ]( i) Z. k# v
Auto: $18 spend, .95% Ad CTR, $12 revenue = -33% ROI
Golf: $75 spend, .35% Ad CTR, $0 revenue = -100% ROI
Skincare: $120 spend, .14% Ad CTR, $40 revenue = -66% ROI! B( R3 e. X% U- }+ Q4 ^
You have to be brutal here. Don't worry about "maybes" or "what ifs" or you'll end up trying to run all 3 campaigns at once. The Auto Insurance campaign had the highest CTR and the best ROI. That's it. 8 Q/ {* Z. E( ?) i9 @ F+ L/ I* ~4 o2 u. m
Pause the 2 lowest performing ad groups. $ d, z9 q( |* Z# v ^5 r, l- D6 v; g1 V4 v
Now we know that we are going to run Auto Insurance because it got the best reaction from the traffic source. This campaign will now be getting a daily budget of 3x offer payout. " t% `. m' b% A1 c8 G 0 g( a/ v+ F& `/ I6 c" SNext - we'll move to the 3-step optimization process, with all steps based on the number "3".! { R" D; _& ?: M9 s& ?. a
5 J% M( z( S- \9 r; A. ^Step 1: Launch Campaign ( t% o* z3 Y* z1 a$ } 9 O" H! G+ Q: J0 F9 T6 w. c( _- E% uIn the first phase of the campaign my goal is to see which ad image and angle are going to get the best reaction.7 _% w" F1 H4 l# e' O- o
) `3 s$ c& |: \; g
Remember, there are 3 kinds of reactions. Ad clicks, LP clicks and conversions.7 @ n& P# x9 e, K9 Y
# S; b$ j+ M/ h3 M/ c, b: A
Note: This first step could definitely result in a big loss of money. It's important to know that before you launch, that way you won't be disappointed and quit after the first 3 days.9 f4 M1 |5 M# ~# C3 F
Create 3 ad groups with 3 ads, each with a different angle
Set the budgets to 1x offer payout per ad group
Run the ad groups for 3 days6 t- v# u4 Y6 T- @, ~% y. N) z+ w
Here's what it looks like: 0 ]0 u, ?, F% i- U + h( p7 P' x8 g5 y* s9 b, dAngles: Cars / Smiles / Money 9 n6 f' h! \/ \( B8 z! V S- k" m7 E0 L" f7 y+ ]7 H0 O0 h0 H
[img]http://madsociety.net/forum/attachments/step2-gif.198/对不起，我们不允许远程图片. 请将图片上传到论坛 , z4 r9 u9 N# F/ J" v+ t7 j+ ?, ` F9 D
Step 2: Choose best ad group / angle 3 [' a* C$ e; f- z0 v ; \5 _: T3 Q! R( KAfter 3 days you will look at which ad group got the best reaction. Look at ad ctrs, landing page ctrs and conversions (if applicable). 1 ^8 v6 o6 N D" d; g' g, r% z8 J2 S& w) H0 h4 ^4 m* O
Conversions will be helpful at this point, but aren't necessary. Even if you have zero conversions just go with the ads that got the best reactions., G& m9 h. E) v1 y) g0 h- o
# P$ _% k: |$ X$ _Normally, one set of ads will "pop" and get a much higher CTR and conversion rate than the others. This set of ads is probably going to be the best route to take.- ~$ @) \7 B3 A) B' C% s
! o- ?- B n, {6 `3 W5 [ Example results after 3 days:7 d: n% ~- i9 U
Ad Group 1 (smiles): CTR .99%, Spend: $18, Revenue: $24 = +33% ROI
Ad Group 2 (cars): CTR .69%, Spend: $18, Revenue: $6 = -66% ROI
In this case the winner is clear. Ads that mention Happiness with people smiling is going to be the best angle to pursue. R3 _1 O8 w7 q8 `9 K L! } f+ ^2 y( c*Ad headlines and ad body copy: If you're profitable at this point, you're done. If not, you'll need to work on your ad headline and body copy. To do this just repeat the exact same process as outline above. . w4 y; _" r- }8 M, b# S2 g7 Y $ r* H) D5 `1 ] F/ x6 s" _Ad Headlines: Write out 3 ad headlines and run for 3 days with a 3x offer payout daily budget. Choose a winner and repeat until you're happy with results. " J$ ]0 A: F5 i+ P; V- \: \4 A' O; b& u6 w* Z Ad Body Copy: Some say that ad body copy is key in getting people to convert. If you're still not profitable work on your ad body copy using the same 3x process. / Z/ L' c9 p/ B7 o7 w( w3 F$ [6 b6 ~1 B& _4 t7 k) I; D% f
P.S. If you want to continue testing and optimizing the "smile angle" you would take the best "smile" ad and make more versions like it. Let's say the "woman smiling" did the best.% N; D3 `) z2 \. N# x& D8 w
, G% T3 m2 N/ n/ M& m) |
You would create 3 versions of woman smiling (here I'm testing ethnicity) and run them for 3 days. 9 i& c# f2 Y% y/ ~, G. g: B3 v: P% F; y* a e5 p6 C1 O3 a0 Q
[img]http://madsociety.net/forum/attachments/step3-gif.199/对不起，我们不允许远程图片. 请将图片上传到论坛 9 b" j: r, u" G
! X% Y7 _% I: B
You can continue doing this. If you find that the asian woman did best, you would now test 3 different asian women smile and doing different things. (driving, walking, jumping)" ~' s& r: W" e }# Y7 Q& C6 K
9 c3 F: n+ I5 }; @9 EStep 3: Scaling# ]4 u$ u# G7 |5 ^7 h( S3 o
0 u' K4 l: Y4 t; V$ d1 U+ r& B4 b" f At this point you should have solid set of ads in one ad group. If you used a proven landing page you should be profitable or close to profitable. . b1 r3 K. |5 \ a3 n: P5 R8 ?# z9 F8 n0 g* A
Now you'll want to start scaling. You've kept the daily budget fairly conservative for the first few weeks to save money while you learn what works, now it's time to scale.0 f2 O: A' J `' G+ X( Z: T5 e, u, B
: q0 l- U( P8 ^6 P. A8 mRaise your daily budget 3x every 3 days. 7 e3 e8 }7 ^' X6 H. z! x9 A* J( [$ J' z5 j( k1 z5 p" O9 [, K
In this example you have an $18/day budget. If after 3 days you want more traffic, raise your budget to $54. 3 days later raise it to $162/day. Feel free to round up or down. - J( { c; `4 P6 a7 P0 n: ^! K5 n% _6 K1 u* U4 ^9 a/ u9 h
If your ROI dips after raising your budget, roll you budget back to where it was before. Then work again on your ads and landing page to get ROI up again, then raise your budget again when you're ready. ' V9 E4 J5 `" L! o6 Y& r4 l, U/ F4 `* j- J" x
This conservative budget scaling protects you from losing money. You don't want to raise your budget from $18/day to $18,000/day. That is a recipe for massive disaster. s) y/ U; S2 \; ?
& [" C8 F4 \: w9 ?Budget Sweet Spot: 6 x' \3 O- S- u# x1 S3 g( h * V8 \" O# x! J \+ WConservative budget scaling may also reveal to you a "sweet spot" daily budget. For example - let's say you raise your budget from $500/day to $1500 and it falls apart. In this case, your campaign may not be able to run at that high of a volume. Some campaigns just can't scale to $10k/day./ f3 F: ?& N; @, V
/ k& C) `, G. p; u) w! o. A1 eRaise your daily budget by 3x every 3 days and keep an eye on your ROI and you'll be safe. If ROI dips, roll the budget back, if ROI is great, raise it up 3x.3 O4 p$ d, K4 H) Y
& t% m, a3 `' e, }& O% O- a; L3 ]
Keep increasing it by 3x until you have a steady flow of profitable traffic that you're happy with. ' r) r7 g4 e6 Z' | 3 i& K% ?- P" B" N8 |9 ?* qCongratulations! You now have a lethal set of ads running and a profitable campaign. 8 S$ N. J6 G( A& [
- Y+ a3 ]3 I- m5 R #1Malan, Feb 29, 2016 X' k0 P1 q7 M l% f9 M$ C3 u. ]+ F" B" p1 a% r+ y2 w
; _8 }8 \7 p. ?' |& R# \( q5 G
[复制链接]
发表于 2016-10-8 16:18:34
发表于 2016-10-8 16:49:17
发表于 2016-10-8 16:56:03
楼主
