Improvement

Introduction

The assignment of marketing costs to transactions is an important feature so that contribution margins including marketing costs can be flexibly calculated from transactional data.

A technical challenge here is dealing with special cases, e.g. “What happens to the marketing costs if there were no transactions for a campaign on a certain date?”. So that the costs are still included in the contribution margin calculation, the cost data must always be anchored in the transactional data. "Nothing should fall under the table."

In the past, there were complex fallback rules to always match all marketing costs to transactions. Example: If there are no conversions for the cost from one source on a day, then the cost will be spread across all conversions with the same campaign type on that date.

The assignment via fallback rules has advantages, but was not very transparent for users and sometimes also led to unfortunate distortions in data analyses. For this reason, we have decided to simplify the allocation of marketing costs and make them more transparent:

Assignable and non-assignable marketing costs

The assigned marketing costs (OD, ID, attributed) are now calculated from two components that are now also visible in the front end:

• Assignable marketing costs is the portion of assigned marketing costs that could be assigned to transactional records.
  • This measure returns results along with transactional attributes (e.g. products, brands, customers,…) and is part of the contribution margin calculation.
• Unassignable marketing costs is the portion of assigned marketing costs that cannot be assigned to transactional records.
  • This is usually the case when there are no touch points from completed customer journeys (attributed) or conversions (OD, ID) on the same date and from the same source.
  • This measure does not return any results together with transactional attributes (e.g. products, brands, customers,…) but is still part of the contribution margin calculation.

The sum of both components is the allocated marketing costs.

Marketing cost assignment process

1. minubo loads the costs per period and campaign from the data source.
2. If the source provides costs for a period longer than 1 day, the costs are converted to daily costs.
3. The costs per day and campaign are written to the Marketing costs measure.

The further steps are only relevant for the 'Assigned marketing costs'.

For 'Associated marketing costs (BD) / (FD)':

4. The daily cost per campaign is distributed across the conversions from this source on this date:

• If there are conversions from the same campaign source on that date, the cost will be distributed across the orders behind those conversions.
  • This is the assignable portion of the marketing costs.
  • Accordingly, this portion of the costs can be evaluated with all transactional attributes.
• If there is no conversion with the same campaign source on that date, the cost will be written to the 'Unassignable Marketing Costs' measure.
  • Accordingly, this portion of the costs can only be evaluated with the sales channel, the marketing channel and the other campaign attributes.

For 'Associated Marketing Costs (Attributed):

4. The daily costs per campaign are distributed to the touch points of completed customer journeys. That is, if there is a customer touchpoint from that source on that date that is part of a completed customer journey, then a portion of the marketing cost is assigned to that touchpoint.

• If there are touchpoints in completed customer journeys with the same campaign source on that date, the cost will be distributed across those touchpoints.
  • This is the assignable portion of the marketing costs.
  • Accordingly, this portion of the costs can be evaluated with all transactional attributes.
• If there are no touchpoints in completed customer journeys with the same campaign source on that date, the costs will be written to the 'Unassignable Marketing Costs'.
  • Accordingly, this portion of the costs can only be evaluated with the sales channel, the marketing channel and the other campaign attributes.

What changes in the data

The change of the assignment process affects the following measures:

• Assigned Marketing Cost (OD) / (ID) / (Attributed)
• Assigned Fixed Marketing Cost (OD) / (ID) / (Attributed)
• Assigned Variable Marketing Cost (OD) / (ID) / (Attributed)
• Contribution Margin III (OD) / (ID,RD)
• Contribution Margin III Rate (OD) / (ID,RD)
• Avg. Contribution Margin III (OD) / (ID,RD)
• Avg. Contribution Margin III per Item (OD) / (ID,RD)
• Avg. Contribution Margin III per Customer (OD)

Specifically, the following effects occur:

• In data queries that connect transactional attributes (e.g. products, brands, customers, …) with one of the above measures, a new row with the value (empty) may appear.
  • The unassignable costs were previously distributed to existing transactions using fallback rules. In consequence, all costs were assigned to an attribute value.
  • These fallback rules no longer exist, therefore the share of marketing costs that could not be assigned to any transaction is not linked to any attribute value.
• The Assigned Marketing Costs per Campaign Level 1-5 change as costs are no longer aggregated via fallback on the 'Campaign > Type' level.

Migration

It was important to us that the names of the existing measures do not change, so that no adjustments to existing feeds and their Excel/ Sheets/ Power BI data are necessary.

Improvement

Einleitung

Die Zuordnung von Marketingkosten an Transaktionen ist ein wichtiges Feature, damit Deckungsbeiträge inklusive Marketingkosten flexibel aus transaktionalen Daten berechnet werden können.

Eine Herausforderung dabei ist der Umgang mit Sonderfällen, z.B. “Was passiert mit den Marketingkosten, wenn es keine Transaktionen für eine Kampagne gab an einem bestimmten Datum?”. Damit die Kosten trotzdem mit in die Deckungsbeitragsrechnung mit einfließen, müssen die Kostendaten immer in den transaktionalen Daten verankert werden. “Nichts darf unter den Tisch fallen.”

In der Vergangenheit gab es komplexe Fallback-Regeln, um immer alle Marketingkosten an Transaktionen zuzuordnen. Beispiel: Falls keine Conversions vorliegen für die Kosten aus einer Quelle an einem Tag, dann werden die Kosten auf alle Conversions verteilt mit dem gleichen Kampagnentyp an diesem Datum.

Die Zuordnung über Fallback-Regeln hat Vorteile, war für Nutzer:Innen jedoch wenig transparent und führte in Teil auch zu unglücklichen Verwerfungen in Datenanalysen. Aus diesem Grund haben wir uns entschieden die Zuordnung der Marketingkosten deutlich zu vereinfachen und transparenter zu machen:

Zuordenbare und nicht-zuordenbare Marketing-Kosten

Die zugeordneten Marketing-Kosten (BD, FD, Attribuiert) berechnen sich neuerdings aus zwei nun auch im Front-end sichtbaren Komponenten:

• Zuordenbare Marketing-Kosten ist der Anteil der zugeordneten Marketing-Kosten, der transaktionalen Datensätzen zugeordnet werden konnte.
  • Diese Kennzahl gibt Ergebnisse zurück zusammen mit transaktionalen Attributen (z.B. Produkte, Marken, Kunden,…) und ist Teil der Deckungsbeitragsrechnung.
• Nicht-zuordenbare Marketing-Kosten ist der Anteil der zugeordneten Marketing-Kosten, der keinen transaktionalen Datensätzen zugeordnet werden kann.
  • Dies ist in der Regel der Fall, wenn am selben Datum und auf der selben Quelle keine Kontaktpunkte von abgeschlossenen Customer Journeys (Attribuiert) oder Conversions (BD, FD) vorliegen.
  • Diese Kennzahl gibt keine Ergebnisse zurück zusammen mit transaktionalen Attributen (z.B. Produkte, Marken, Kunden,…), aber ist dennoch Teil der Deckungsbeitragsrechnung.

Die Summe aus beiden Komponenten sind die zugeordneten Marketing-Kosten.

Marketingkosten Zuordnungsprozess

1. minubo lädt die Kosten pro Zeitraum und Kampagne aus der Datenquelle
2. Falls die Quelle Kosten für einen Zeitraum liefert, der länger ist als 1 Tag, werden die Kosten zu täglichen Kosten umgerechnet.
3. Die Kosten pro Tag und Kampagne werden in die Kennzahl Marketingkosten geschrieben.

Die weiteren Schritte sind nur für die ‘Zugeordneten Marketingkosten’ relevant. Für 'Zugeordnete Marketingkosten (BD) / (FD)':

4. Die täglichen Kosten pro Kampagne werden auf die Conversions aus dieser Quelle an diesem Datum verteilt:

• Falls Conversions mit der gleichen Kampagnenquelle an diesem Datum vorliegen, werden die Kosten auf die Bestellungen hinter diesen Conversions verteilt.
  • Dies ist der zuordenbare Anteil der Marketingkosten und entsprechend in der Kennzahl ‘Zuordenbare Marketingkosten’ zu finden.
  • Entsprechend ist dieser Anteil der Kosten mit allen transaktionalen Attributen auswertbar.
• Falls keine Conversion mit der gleichen Kampagnenquelle an diesem Datum vorliegen, werden die Kosten in die Kennzahl ‘Nicht Zuordenbare Marketingkosten’ geschrieben.
  • Entsprechend ist dieser Anteil der Kosten nur mit dem Verkaufskanal, dem Marketingkanal und den übrigen Kampagnen-Attributen auswertbar.

Für 'Zugeordnete Marketingkosten (Attribuiert):

4. Die täglichen Kosten pro Kampagne werden auf Kontaktpunkt von abgeschlossenen Customer Journeys verteilt. Das heißt, wenn es einen Kundenkontaktpunkt aus dieser Quelle an diesem Datum gibt, der Teil von einer abgeschlossenen Customer Journey ist, dann wird diesem Kontaktpunkt ein Teil der Marketingkosten zugeordnet.

• Falls Kontaktpunkt in abgeschlossenen Customer Journeys mit der gleichen Kampagnenquelle an diesem Datum vorliegen, werden die Kosten auf diese Kontaktpunkte verteilt.
  • Dies ist der zuordenbare Anteil der Marketingkosten und entsprechend in der Kennzahl ‘Zuordenbare Marketingkosten’ zu finden.
  • Entsprechend ist dieser Anteil der Kosten mit allen transaktionalen Attributen auswertbar.
• Falls keine Kontaktpunkt in abgeschlossenen Customer Journeys mit der gleichen Kampagnenquelle an diesem Datum vorliegen, werden die Kosten in die Kennzahl ‘Nicht Zuordenbare Marketingkosten’ geschrieben.
  • Entsprechend ist dieser Anteil der Kosten nur mit dem Verkaufskanal, dem Marketingkanal und den übrigen Kampagnen-Attributen auswertbar.

Was sich in den Daten ändert

Die Änderung des Zuordnungsprozess wirkt sich auf folgende bestehende Kennzahlen aus:

• Zugeordnete Marketingkosten (BD) / (FD) / (Attribuiert)
• Zugeordnete variable Marketingkosten (BD) / (FD) / (Attribuiert)
• Zugeordnete fixe Marketingkosten (BD) / (FD) / (Attribuiert)
• Deckungsbeitrag III (BD) / (FD,RD)
• Deckungsbeitragsquote III (BD) / (FD,RD)
• ø Deckungsbeitrag III (BD) / (FD,RD)
• ø Deckungsbeitrag III pro Artikel (BD) / (FD,RD)
• ø Deckungsbeitrag III pro Kunde (BD)

Konkret kommt es zu folgenden Effekten:

• In Datenabfragen, die transaktionale Attribute (z.B. Produkte, Marken, Kunden, …) mit einer der genannten Kennzahlen verbinden, wird ggf. eine neue Zeile mit dem Wert (leer) auftauchen:
  • Die nicht-zuordenbaren Kosten wurden zuvor über Fallback-Regeln auf bestehende Transaktionen verteilt, sodass alle Kosten immer einem Attribut-Wert zugeordnet waren.
  • Diese Fallback-Regeln gibt es nun nicht mehr, dass heißt, falls ein Anteil an Marketingkosten nicht zuordenbar ist, wird dieser Anteil der Kosten auch keinem Wert im betrachteten Attribut zugeordnet sein.
• Die Zugeordnete Marketingkosten (BD) / (FD) / (Attribuiert) pro Kampagne Level 1-5 ändern sich, da keine Kosten mehr per Fallback auf der Ebene 'Kampagne > Typ' aggregiert werden.

Migration

Es war uns wichtig, dass sich die Namen der bereits bestehenden Kennzahlen nicht ändern, sodass keine Anpassungen von bestehenden Feeds und deren Excel/ Sheets/ Power BI Daten notwendig ist.

New

Long anticipated, now here: The restructured minubo Suite is here.

Until now, the minubo suite was structured according to its tools, with a separate page for each tool, an explorer for the tool's entities and a central navigation bar in the application header.

• The tool-based structure of the application has been replaced with a common explorer for all dashboards, queries, reports and analyses. That means you can find all existing dashboards, reports, queries and analyses in the explorer on the left side. This will make it much easier for you to organize all your files in minubo, since you no longer need to create and maintain similar folder structures multiple times.

• You create new dashboards, queries, reports and analyses from one central location, the "New" button. You find all minubo tools in one place!

• With the new search you can quickly find all dashboards, reports, queries, analyses and folders. Cool: You can also search for attributes and measures to call up their documentation.

• You can mark your favorite dashboards, reports and queries directly as favorites.
• Recently and frequently used entities are displayed for quick access directly above the explorer.

• The "old" tools (Feeds ("Legacy Feeds"), Webpivot ("Legacy Analysis"), Heartbeats ("Legacy Heartbeats") and Customer Segmentation ("Legacy Customer Segmentation")) can be found in the menu on the left under the item “Legacy Tools”.

• The start page shows the previous functional roles in the new guise. These will be called "KPI sets" in the future. The aim of the new KPI sets will be that all users can find their key performance indicators directly on their start page. For this purpose, the KPI sets have been revised so that each key figure can be filtered separately.

New

Lange angekündigt, jetzt ist es soweit: Die neu strukturierte minubo Suite ist da.

Bislang war die minubo Suite nach ihren Werkzeugen strukturiert mit jeweils einer eigenen Seite für jedes Tool, einen Explorer für die Entitäten des Tools und eine zentrale Navigationsleiste im Kopf der Applikation.

• Die Werkzeug-basierte Struktur der Applikation wurde durch einen gemeinsamen Explorer für alle Dashboards, Queries, Reports und Analysen ersetzt. Das heißt, Du findest alle bestehenden Dashboards, Reports, Queries und Analysen im Explorer auf der linken Seite. Du wirst hierdurch alle Deine Dateien in minubo deutlich leichter organisieren können, da man sich ähnelnde Ordnerstrukturen nicht mehr mehrfach anlegen und pflegen braucht.

• Du erstellst neue Dashboards, Queries, Reports und Analysen von einem zentralen Ort aus, der “Neu” Schaltfläche. Du findest also zukünftig alle Werkzeuge von minubo an einer Stelle!

• Über die neue Suche findest du schnell alle Dashboards, Reports, Queries, Analysen und Ordner. Cool: Du kannst auch Attribute und Kennzahlen suchen und darüber deren Dokumentation aufrufen.

• Deine favorisierten Dashboards, Reports und Queries kannst Du direkt als Favoriten markieren.
• Kürzlich und häufig verwendete Entitäten werden Dir direkt oberhalb des Explorers im Schnellzugriff angezeigt.

• Die "alten" Tools (Feeds ("Legacy Feeds"), Webpivot (“Legacy Analysen”), Heartbeats (“Legacy Heartbeats”) und Kundensegmentierung (“Legacy Kundensegmentierung”)) findest Du im Menü auf der linken Seite unter dem Punkt “Legacy Tools”.

• Die Startseite zeigt die bisherigen Funktionalen Rollen im neuen Gewand. Diese heißen in Zukunft sprechender "KPI Sets". Das Ziel der neuen KPI Sets wird es sein, dass all Nutzenden Ihre Key Performance Indikatoren direkt auf Ihrer Startseite finden. Dafür wurden die KPI Sets dahingehend überarbeitet, dass jede Kennzahl für sich separat gefiltert werden kann.

Minubo Suite / Now / Store Monitor

• We are in the last steps fixing and tweaking before releasing our new analysis tools.
  • The release allows users to perform Cohort, Histogram and Grid analysis within minubo.
  • As soon as we are ready we will give a more detailed update.
  • Meanwhile, we have started the development of a cross-sell analysis tool which later will allow to understand which products, brands, categories are bought together in the same order or also over multiple orders.
• In the background, we are preparing the migration of legacy feeds and web pivots to the current Queries and Feeds tools.

Data Model

• As a custom development, we added a nearest store feature for one of our clients.
  • A custom data pipeline calculates for each customer his/her nearest store and the distance.
  • The information is available within Minubo for analysis and more importantly for targeting specific customer segments!

Data Sources & ETL

• We have added support for the Article Tree from Xentral.

Tech

• We are preparing an upgrade to a new Elastic Search version.

Minubo Suite / Now / Store Monitor

Cohort Analysis

• Work continued on the cohort analysis.
• We expect to finish development in around 2-3 weeks and prepare the release.

Histogram Analysis

• We gathered a first round of feedback last month and are busy implementing it.
• We expect to finish development in around 3-4 weeks.

Grid Analysis

• A first version of the grid analysis is live on our test systems and we are gathering feedback internally and are fixing early development bugs.

Data Model

• Added new attribute "Purchase > Oder > Type" to differentiate between different types of purchase orders (NOS, Nachorder,…)
• Added new attribute "Sale > Order Date > Hour" to analyse how order performance varies over the hours of the day
• Added new measure "Max. Order Processing Time (Hours)" to identify the longest order processing time per channel, store, delivery method, …

Data Sources & ETL

• Xentral's "Hersteller Name" as new source for "Product > Brand"
• Xentral's "Kunde Gruppe" as new source for "Customer > Group"
• Added support for Smart Display Campaigns.

1. Minubo Suite / Now / Store Monitor

New

Histogram Analysis

• We are in the process of adding a new analysis tool: Histograms!
• Histograms help understand the distribution of a measure visually by grouping measure values in bands of equal size. Typical cases come as predefined entities. Among them…
  • What is the ø Order Value per Order?
  • What is the Discount Rate per Order?
  • What is the ø Discounted Merchandise Value per Item per Product?
• We expect the Histogram tool to be released within Q1 2022.
• Additionally, we are in the process of turning the old grid reports into more configurable grid analysis, stay tuned.

2. Data Model

• Added support for more product detail view related rate measures: Adds and Checkouts per Product Detail View. These measures make it easier to identify which products are often added to baskets but rarely bought.
• Added support for aggregated E-mail Touchpoints. This data allows basic statistics like sends, opens and clicks per date and campaign without the overhead (in terms of time and costs) of processing recipient level data.

3. ETL & Infrastructure

• We added support for Amazon Sponsored Product Adspend
• The shopify interface now supports multiple shops.
• We have optimized the data load from Google Analytics

1. minubo Suite / NOW

Cohort Analysis

• We are in the process of developing a cohort analysis tool
• The cohort analysis tools allow you to quickly understand customer retention and the lifetime value of a cohort.
• Unlike other static analysis tools, our cohort tool allows users to analyze cohorts very flexibly with their own parameters for time, intervals, measures and filters.
• In particular, it is possible to apply filters on acquisition and follow-up orders independently allowing for more specific cohort definitions, e.g. Only customers acquired through SEA campaigns.
• Typical views come as predefined entities:
  • Retention View: How many users are loyal customers after n periods of time? Which cohorts are more loyal than others?
  • Customer Lifetime Value View: What is the lifetime value of a cohort and after how many periods does the acquisition pay off?
• We plan to drop this new analysis within Q1 2022. Stay tuned for more.
• We applied our new color scheme to the minubo application
• Added a basic start page to minubo Now

2. Data Model

• Added support for Google Analytics’ Product Detail views and related success indicators, Cart to Detail rate and Buy to Detail rate

3. Data Sources, Interfaces and Data Transformation

• Anonymous customer journeys can be calculated across different touch points while maintaining full protection of data privacy.
• Amazon Advertisement Interface is in the making
• Optimization of real-time processing

1. minubo Suite / NOW

Login Security

• Added a more restrictive standard for password security
• New passwords must comply with password security rules
• Users with weak passwords are suggested to change their passwords

Embedded Dashboards v3

• Build the technical groundwork to allow embedding v3 dashboards

Now Start Page

• Now gets a start page with last used, favorite and most used entities
• Additionally, a special version of our functional roles page is in the making
• Favorites, last used and most used will become available features for the Suite as soon as we replace the Suite Start Page sometime in 2022

Other

• Unified data model pages in the settings area

2. Data Model

No changes

3. Data Sources, Interfaces and Data Transformation

• Features to reduce the data history of a tenant while keeping customer relationship measures and attributes correct (e.g. number of new customers and acquisition attributes based on full history but only partial history integrated into data warehouse)
• Changed mapping of product numbers and bundles for Pixi
• Skipping successful tasks in retry processes reduces time to reprocess a tenant
• Initial stock calculation is now calculated for the first date in the data history instead of being allocated on the last day of inventory movement of one SKU. This change increases historical inventory values and quantities in case that a max data history is defined for your tenant.

Facebook

• Removed Facebook UTM mapping fall backs. In case a campaign has badly configured UTM parameters, a tenant can now choose to have matching data with Google Analytics (not set) or fields be filled via fall backs from minubo.

Xentral

• Added article types
• Added improvement to fee products
• Xentral inventory movements include warehouse name

4. Bug

• Fixed an issue causing Facebook ad costs to be incomplete
• Fixed an issue causing Facebook ad creatives to not be updated
• Fixed an issue causing the measure “Time since first order (days) (sku level)” to not be filled anymore
• Removed duplicate lines in SQL reports from Xentral
• Fixed an issue causing the feed cache to not respond
• Fixed an issue causing the measures to disappear when switching from Pivot to Table visualization
• Fixed an issue with missing period labels
• Fixed an issue with PDF reports not containing comparison data
• Fixed an issue with the Feeds Overlay not displaying recipients when only external recipients

1. minubo Suite / NOW

Many smaller issues:

• A user can configure credentials for PIXI, Google Analytics, Facebook via settings pages.
  • Suite users require the Data Source Manager right.
• A user can configure the source fields from PIXI for channel level attributes via the Front-end.
• A user can adjust default drill-downs via the data model settings pages. Page allows adding custom drill-down paths as well.
• A user can configure alias for measures and attributes via a data model settings page.
• A user can add comparison periods to queries.
• CSV-file names include query name, feed name and date to be better identifiable
• It is now easier to configure the Facebook and Google Analytics interface pages

2. Data Model

• Added Customer > Account > Type / Kunde > Konto > Typ for B2B / B2C distinction
• Added Customer > Alternative Number for secondary identification numbers

3. Data Sources, Interfaces and Data Transformation

• Xentral interface improved
  • Added Stornorechnungen / Gutschriften
  • Added Bundles handling to analyse sales for bundles and their components at the same time
  • Added order-split
  • Added Product Categories
• Enrichment of voucher codes from Shopify

4. Bug

A bug hunt this month:

• Specific query can lead to 500 Error when downloaded
• Wrong time changes detected on custom dates
• Delete button on feed view page with error
• Credit Notes were missing translation
• Incomplete filters resulting in empty feed in special cases
• Static feeds export is started directly after transformation
• Customers with several orders do not have Sale > Customer > Type in rare cases
• Order lines split (multiple InvoiceLines for one OrderLine) were not supported
• Cancelled quantity not correct