Skip to content
Two people test a headset in an office, seen through glass
  1. Conglomerado
  2. Calaa Mind
  3. Marketing
  4. Cal-1, the app

02.1

The digital product, and the only one that takes advertising.

An app is won on the third open

Budget is allocated against the third open, not against the install. That decides which channel is switched on and in what order.

02.1.1

The download gets published and the third open decides

What the plan pursues, and what the budget is allocated against.

An automotive app gets installed when there is a procedure ahead, whether a purchase, a service or a sale, and it leaves the screen as soon as that procedure ends. Installs therefore measure the market season above all. The number that says whether the product exists is how many users come back with no procedure involved.

A campaign that multiplies installs without moving that number has bought traffic, and traffic does not stay. The third open is the stated objective and everything else is ordered behind it.

The app belongs to the group, and that governs the whole plan. When measurement shows that drop-off concentrates on the second step of sign-up, the answer is to rebuild that step, not to raise media spend to compensate.

HOW FAR EACH PUSH REACHES

Each push runs out before the next one starts, and between the second open and the third there is none

It ranks, it does not measure. What each bar states is how far that push reaches, not how many people it moves.

Install
1st open
2nd
3rd
4th
5th
nothing pushes here
the third open

Tap a bar to withdraw that push. The screens that depended on it alone go dark. Tap a screen to see what switches it on.

Advertising cannot buy the third open. That is why budget is judged against it and not against the install.

02.1.2

The five channels

Five channels, each with what it solves and what it costs. They are not switched on at once.

  1. Search

    People search for the problem, never for the name of the app.

    Someone facing a breakdown or a procedure types exactly that. It is the cheapest channel per contact and the slowest to deliver, because what gets published takes weeks to be found and then works for years.

  2. The store listing

    Fixing a listing consumes no media budget.

    Screenshots, the first line and the ratings decide more installs than the money that drives traffic there. It is the first thing reviewed when cost per install rises.

  3. Short video

    It turns curiosity into installs better than any other channel and withstands repetition worst.

    An industry procedure explained in thirty seconds, such as what to check on a spec sheet or what paperwork a transfer of ownership requires, travels with no spend behind it. The format wears out as soon as it repeats, so it demands new production continuously.

  4. The workshop and the dealership

    The only channel with a person in front of you.

    An install completed at the counter, next to the customer, stays longer than one won through advertising. In exchange it requires training and a reason for whoever offers it, and that is paid in time rather than in media.

  5. Paid advertising

    It accelerates what already works and is no use for finding out whether something works.

    It is switched on when an owned channel gives a stable number, and switched off when cost per useful contact rises above what that contact returns. Switching it on before that reading exists means paying to repeat a mistake faster.

A screen showing a long run of market data

02.102.1 · The physical channel

A counter leaves permanence. An advert leaves an install.

A workshop that shows the app at the counter leaves a user who comes back. The same install won through advertising lasts as long as the procedure that prompted it.

02.1.3

The order of the four steps

Four steps, and the order weighs more than any of them on its own.

  1. 01

    Retention is fixed before traffic is bought

    Step 1

    Driving traffic to a product that does not retain widens the reach of the fault. The first job is the screen where drop-off concentrates, and it consumes no media budget.

  2. 02

    The starting point is measured

    Step 2

    Installs, opens and completed procedures, dated, before the first action. Without that prior value there is no way to separate the effect of a campaign from the effect of the market's high season.

  3. 03

    One channel is opened at a time

    Step 3

    Until it gives a stable reading over several weeks. With three switched on at once, the combined result cannot be split between them and the next decision is taken blind.

  4. 04

    Budget is allocated against what it returns

    Step 4

    Each week, budget moves to the channel whose contact comes cheapest and reaches the third open. Whatever fails to reach that third open stops receiving money even if it brings volume.

WHAT EACH SKIPPED STEP DRAGS WITH IT

The cost of skipping a step is decided by its position in the order

It ranks, it does not measure. The number under each step states how many are left unjudgeable if that one is missing, not what it costs.

01Retention is fixed before traffic is bought3 left unjudgeable
02The starting point is measured2 left unjudgeable
03One channel is opened at a time1 left unjudgeable
04Budget is allocated against what it returns

Tap a step to skip it. The chain breaks there and what came behind goes dark. Under each one is how many it drags.

The order is not a preference of method. A step that gets skipped does not lose only itself, it takes everything behind it with it, and the first one takes all three.

02.1.4

What gets reviewed every week

Four numbers that decide, and one that is recorded without being pursued.

  1. The third open

    It is the only number on the list that cannot be bought.

    How many users open the app a third time without a procedure prompting it. Budget buys installs and opens; the third one is decided by the product.

  2. Cost per useful contact

    Cost per click measures curiosity above all.

    This counts what it costs to reach someone who fits the segment and goes on to use the app. The two numbers are nothing alike, and the one that governs allocation is the second.

  3. Time to first procedure

    When it rises, no campaign fixes it.

    How long a user takes from installing to completing a procedure. It is the number that warns the problem has left the channel and moved to the product.

  4. Drop-off by screen

    It turns a marketing problem into a product task.

    Which step loses the user. It exists because the app belongs to the group, and so the answer can be a product fix instead of a media purchase.

  5. Installs and followers

    They are recorded and not pursued.

    Installs and followers explain the cost of the other numbers, so they are reviewed. Setting them as the objective is what leads to buying audience at quarter end.

02.1.5

Both numbers at once, or it does not count

The third open rises while cost per useful contact falls or holds, and both at once. Either one on its own is achieved by switching off the other, because cost falls by buying cheap traffic that never returns and permanence rises by giving up on acquisition. A good quarter where nobody knows which of the two produced it cannot be repeated the next one.

THE TWO TRAPS OF MEASURING A SINGLE THING

Either of the two numbers can be improved on its own, and always at the expense of the other

It ranks, it does not measure. Each arrow marks the direction that number moves in, not how far it moves.

The third open, upwards
Cost per useful contact, downwards
Acquisition has been given up
Both at once
Nothing to discuss
Cheap traffic that never returns

Two numbers, four readings. Each arrow states where that number goes in that case. Three columns carry an arrow on the wrong side; the one that remains carries none.

A good quarter where nobody knows which of the two produced it cannot be repeated the next one. That is why the plan is judged on both at once, and of the four combinations only one holds.