Google Ads Strategy When Taking Over a Failing Campaign Part 1 - The Google Ads Podcast

Artwork for podcast The Google Ads Podcast

Google Ads Strategy When Taking Over a Failing Campaign Part 1

23rd July 2024 • The Google Ads Podcast • Solutions 8

John Moran: 00:00:10

I can't generate demand.

2

That's not what PMAX does best.

3

So it's simply just attributes.

4

It's not going to be able to attribute

as high of a row as it did before.

5

This actually has nothing to

do with actual backend metrics.

6

This is just the audience.

7

It can go after how valuable was it?

8

How warm was it?

9

And where did it come from?

10

We all don't have any answers

to that just as of yet.

11

When we take over a campaign,

two things will take place.

12

One it is previously probably used to

remarketing, especially with, and we

13

don't know the levels though of meta.

14

What is, we have access to GA4?

15

Yep, we do.

16

What is the volumes between social and

paid, search look like for last 30 days?

17

What I'm going to slightly assume

is that there is potentially

18

a pay, social traffic as well.

19

All right.

20

So we have a three 90.

21

the highest is from

organic and cross network.

22

It's not showing anything about social.

23

Let's do user acquisition.

24

All We don't know if they're

using any sort of UTMs,

25

though, inside of paid social.

26

Yeah, we'll need to see

what he's doing on meta,

27

at least just from a spend level.

28

because what this is saying, he's not

doing any meta, which I don't believe is

29

correct, because it could be using landing

pages, that could be doing something

30

else that we just don't see here.

31

that's part two, where we'd like to get a

little bit more, a little more data there.

32

Moving back over to the Google ads point.

33

Okay.

34

Yeah.

35

a couple of things, this campaign here,

is on a steady decline, but they're

36

fairly equal of costs and conversions.

37

This is since January.

38

Now, one thing to note is the conversion

value by cost has fallen below a three

39

90, right after we took over here at

two 67 and two 80, but this I would

40

imagine is there, there's a trend here.

41

That's bigger than what we can see.

42

this one is going to be a downtrend.

43

In terms of, is it a seasonality

B R T row as targets have always

44

been a little bit too high.

45

This right now at three 90, when we're

getting a two 60 and two 70 is just,

46

is going to cost some underspending.

47

So if we look at like the last seven

days and we segment by day, we can see

48

that with a T row as, and I actually just

put a target here, a target row as this

49

will tell us what that target row as a

set to, and you'll see that there's a.

50

This is okay.

51

Get three 90.

52

And so it's miss.

53

We got it here.

54

but because the average is two 80,

it's going to say, the only way to get

55

up to three 90 is by spending less.

56

The problem though, is if this

is not doing actual cold traffic,

57

this thing is just going to react

to demand, not generate demand.

58

And just drag itself down with it.

59

this one here is at a also three 90.

60

and there would be remaining products.

61

Sometimes it will hit it.

62

Sometimes it will not.

63

but because the average

on last seven days, 3.

64

24, where a time lag is two days.

65

it's going to cause also an additional

underspend, but what we don't know

66

is what was PMX doing before we know

that it had a bunch of remarketing

67

audiences and we know that it can

take credit for any touch point.

68

And because this may have.

69

Taken all of our warm traffic

and simply just removed it.

70

It's not going to be able to attribute

as high of a row as it did before.

71

This actually has nothing to

do with actual backend metrics.

72

This is just the audience that

can go after how valuable was it?

73

How warm was it?

74

Where did it come from?

75

We all don't have any answers

to that just as of yet.

76

So this is essentially saying

take all the warm traffic out,

77

but still get me what you used to.

78

It says, I can't generate demand.

79

That's not what PMAX does best.

80

So it's simply just at atrophies.

81

usually I would say drop that T

ROAS goal, but because the client

82

is only going to measure us by

in app T ROAS, can do two things.

83

One, have long, hard conversations

with the client by saying, here's

84

why you hired us and what we can do

and how we can grow your business.

85

Or two, acquiesce to his requests

and simply throw in brand and

86

remarketing audiences, existing

customers, and just fluff that up

87

until his business maybe stabilizes or.

88

Dies, but it's going to

be heavily on that too.

89

Yeah.

90

he'd be totally willing to have

that, do this or things go down

91

the tubes real fast conversation.

92

Cause he's he, I just didn't

like, I was covering for any while

93

she was gone with family stuff.

94

And he basically just emailed us

Friday or Saturday and said that.

95

he's already in the position where they

have to make a decision about whether or

96

not they want to even keep going or not.

97

So it's better to have

the hard conversation now.

98

Yeah.

99

Yeah.

100

And I think we don't have any sort of

metrics on CAC or all the other stuff.

101

that's where we'll definitely want

to say, Hey, that's the one thing

102

to know with everyone is P max not

always rarely, generates demand.

103

it will capture.

104

very well.

105

Will it generate demand?

106

Not as often.

107

It's just a good rule of thumb to think

about when you're developing a strategy

108

is will this generate demand or will

it react to demand on a downtrend of a

109

business and in a campaign that won't

react or won't generate demand that's

110

where you start to see that kind of like

pushing rope thing where it's just going

111

to slowly decay from 100 down to 30.

112

As the client's volume dies off, at these

levels here, it's hard to pay our own fee.

113

So those are the thing is, always just

keep that in the back of your mind,

114

which is, are we able to generate demand?

115

Not maybe necessarily

the standard shopping.

116

We only spend a hundred bucks.

117

in the last week out of 1000.

118

So we may be really, we have 95 cents.

119

So it's okay.

120

like these aren't really doing much.

121

so we do have standard shopping, but

I think by measuring in app, that's

122

where we chase our tail into submission.

123

what I would say is.

124

In order for us to come back

to him and say, Hey, we have

125

to have the hard conversations.

126

I would definitely point

to this trend here.

127

When there is cost, there is conversions.

128

But since January, you've been

on a slow and steady decay.

129

If we're looking at chasing

ROAS into the ground.

130

and chasing ROAS into the ground

means that, even expand this one here

131

and be like, that is not gonna help.

132

That chasing row has is where we will

win the battle and you'll lose the war.

133

So you'll pay us until

you can't pay us anymore.

134

so that's always that kind of hard

conversations that we would end up in,

135

but, what we say is looking at, new lead

generation and new client acquisition.

136

we had to flip this graph.

137

We have to start to spend up,

restricting it by don't spend unless

138

you can does not generate demand.

139

that is reacting to demand and

that's true across every platform.

140

if you don't spend unless you generate

demand, it's like a car before the horse.

141

It's Hey, go tell all those

people about this new product.

142

And they're like, okay, but don't

tell anybody that's not interested.

143

How how do I introduce with restrictions?

144

So it's like an oxymoron for Google.

145

So that means that it'll just

it'll head off into the, there

146

is not a Forex here, so I will

just stop spending kind of thing.

147

the detailed part of this is we

don't necessarily know how this has

148

been trained since, the inception.

149

which means if we look at all time.

150

These campaigns have been

trained since June of: 2022

151

this will take, approximately 90

days to quote unquote relearn.

152

And they may never really relearn

if we keep the restriction here by

153

saying, don't learn, So it's yeah,

so that's, and again, this isn't a,

154

this is me, not me beating it up.

155

it was me just saying here's some

good ways to think about this with

156

when we go to the client and say,

Hey, Nfroez is cool, but, we're

157

using omni channel campaigns in a.

158

A known omni channel marketing

platform, which is your business.

159

And we don't know where

these are coming from.

160

We know we can generate demand, but if

this actually can head in the really

161

right direction, if they spike meta or

if meta does better, or if meta is doing

162

well, this will start to react to that and

it'll look like a really good high rise.

163

And we've all been privy to that when

people are like, Oh, we spend five

164

times more than us on meta and they're

increasing spend is doing really well.

165

And we just ride on those coattails.

166

That's what PMX will normally do.

167

but yeah, this is just a good,

I'm glad you brought this one up.

168

Thank you.

169

Ready?

170

I know this is you're like in the

hot seat, but not really, cause

171

this is never anything personal.

172

It's just more of how to leverage

Google's, algorithms, I guess I would say.

173

yeah, the insights reports look good.

174

the, good non brand cold traffic.

175

This may actually work really well.

176

If you just switch this over to

maximize conversion value, honestly.

177

This could be as simple as that.

178

the only reason why this is not

necessarily performing is just because

179

there are two rows restrictions

are too high in a channel.

180

or in a company known as

Google that doesn't attribute,

181

completely well, Glenn, what's up?

182

And, I was actually just thinking of

that just before you said it, what about

183

changing the, the max conversion value?

184

Yeah.

185

Then set up a search for shopping.

186

YouTube display RLSA, and that

handles all the retargeting.

187

That's the only Let the PMAC do the

prospecting, theoretically, with no T

188

ROAS, and then set up RLSA other channels.

189

And that's exactly right.

190

The first thing I honestly would do is

just take off the T Row as, see if this

191

can generate any demand on its own.

192

If it cannot, because even though,

yes, there is no remarketing signals

193

in there, it will still do remarketing.

194

that's something I've proven so many

times with my own campaigns where I'm

195

like, all right, I'm now suggesting

things that Google is just ignoring.

196

And then, and sometimes for good reason.

197

but I think what's interesting

about this is here's the good news.

198

Is the blue line and the

red line are matched.

199

And that's where I see light at the

end of this tunnel is if the blue line

200

was like this and the red line was like

that, Oh, but this could be that, if you

201

flip this graph around and we took off

the T row as, and the blue line and the

202

red line did this, then problem solved.

203

Because search terms that

we can see look decent.

204

Like they don't look bad.

205

That's what I'm saying is this is not

something I think that is, it's super

206

specific, like to a really crazy degree,

which is why I think TRO as is doing what

207

it's doing, it's ha, I found somebody

and then it converts them, we're really

208

not pushing in areas that were unknown.

209

So it's like, all right, we got one

conversion from Legrand tango violin.

210

How many clicks did we get on that one?

211

One.

212

that's what t roas is doing.

213

It's I found a person, and

then, boop, converts them.

214

Which, okay, good.

215

It knows people.

216

How many doesn't know?

217

We have to find out by

removing that t roas.

218

And that's when I see this thing

is not that's what I'm saying.

219

It's not generating demand.

220

It is reacting to demand with less than 10

clicks and they in any sort of direction.