Un processo
trasparente
e garantito
basato su tecnologie
d’avanguardia

A wide social survey, a rigorous method

La Ricerca di Base Auditel è una grande indagine sociale che: 1. sette volte l’anno, interroga – esclusivamente nella loro abitazione – 20 mila famiglie residenti in Italia; 2. alimenta una grande banca dati; 3. permette di estrarre il campione destinato a formare il SuperPanel™ Auditel. La Ricerca di Base Auditel consente di scattare una fotografia in alta definizione del fenomeno televisivo in Italia grazie alla dimensione del campione (41 mila individui); alla dispersione del campione (oltre 3.500 degli 8.000 comuni italiani); all’approfondimento delle informazioni raccolte; e alla metodologia usata, ossia Computer Assisted Personal Interview (CAPI) face-to-face. Applicando rigorosi criteri statistici, Auditel ha costruito, attraverso la Ricerca di Base, un campione (il SuperPanel™) altamente rappresentativo della popolazione italiana su basi geografiche, demografiche e socioculturali: tutti gli individui d’età superiore ai 4 anni (dati Istat) residenti sul territorio nazionale. Tanto che la Ricerca di Base – grazie anche alla preziosa collaborazione con Banca d’Italia e con ISTAT – è ora profondamente incardinata nel sistema statistico nazionale, giacché permette di scattare una fotografia unica del Paese e delle sue trasformazioni. La Ricerca di Base Auditel è il risultato di una serie continuativa di indagini generali sulle famiglie residenti in Italia che alimentano, scandite da 7 rilevazioni annuali, una grande banca dati. Partita dopo una vasta ricognizione su 41 mila casi, la Ricerca di Base intervista, ogni anno, nella loro abitazione (e non al telefono), 20 mila famiglie residenti in Italia per stimare, tra l’altro, la dotazione di TV, Smart TV, PC, Tablet e di tutti i device di intrattenimento (quali Set Top Box, VCR, lettori DVD, Game Console). L’intervista domiciliare è l’unica che consente una stima non distorta di fenomeni di penetrazione complessi da rilevare: o perché necessitano di particolari conoscenze tecniche; o in quanto richiedono di accedere agli apparecchi presenti in famiglia. I questionari sono uno strumento indispensabile per la selezione del campione e la sua successiva ponderazione. Infatti, le caratteristiche strutturali della popolazione rappresentata da Auditel vengono in parte ricavate dalle informazioni fornite dall’ISTAT, in parte stimate attraverso la Ricerca di Base Auditel. Dal serbatoio della Ricerca di Base viene estratto il campione destinato al panel ‘meterizzato’, ossia il SuperPanel™ Auditel, utilizzato per rilevare gli ascolti televisivi. Le famiglie della Ricerca di Base vengono selezionate a partire da un complesso sistema geografico che viene costruito ad hoc in modo tale che la probabilità di campionamento sia uguale per tutte le famiglie, sia italiane che straniere, in funzione della zona di residenza effettiva. Questa indagine viene utilizzate per due scopi: 1. determinare l’Universo delle Famiglie TV e degli individui dai quattro anni in su (4+); 2. alimentare una grande banca dati da cui si estraggono le famiglie che verranno reclutate per formare il campione, ossia il SuperPanel™. Grazie alla Ricerca di Base, gli Universi delle Famiglie TV possono essere accuratamente quantificati e dettagliati in termini di: 1. zona di residenza (regione, provincia, city size); 2. caratteristiche sociodemografiche (numero di componenti della famiglia, età, attività, condizione socio-economica); 3. disponibilità di varie apparecchiature televisive e di intrattenimento familiare (televisori, set top box, videoregistratori, lettori DVD, console di giochi etc.); 4. accesso alle diverse piattaforme di ricezione TV; 5. disponibilità di PC, Tablet e SmartPhone e loro appartenenza ai singoli membri della Famiglia TV. Le famiglie incluse nel SuperPanel™ sono estratte e reclutate con metodi che ne garantiscono l’anonimato e la reale rappresentatività rispetto alla popolazione legalmente residente in Italia. Il campione che ne deriva ha una composizione ed un sistema di ponderazione tali da rappresentare correttamente il collettivo di riferimento, più ampio, ossia la popolazione italiana, da cui sono state estratte. La rappresentatività si basa su: 1. un sistema di celle di reclutamento nelle quali si incrociano variabili geografiche (aree e dimensioni dei comuni), caratteristiche strutturali delle famiglie (età del capofamiglia e numero di componenti), dotazione di apparecchiature televisive; 2. un doppio sistema di espansione (uno per i dati familiari e uno per quelli individuali: l’espansione è una procedura che consente di trasformare dati campionari in stime riferite alla popolazione rappresentata) caratterizzato da una pre-espansione per celle e da una successiva pesatura iterativa marginale. Il campione è distribuito nelle 103 province italiane in modo proporzionale alla popolazione. La dispersione territoriale del panel consente di coprire oltre 3.500 degli 8.000 comuni italiani. La Ricerca di Base rappresenta, infine, un preziosissimo strumento di indagine sociale. Da essa è nato il ‘Rapporto Auditel-Censis su convivenze, relazioni e stili di vita delle famiglie in Italia’, straordinario racconto della vita vera delle famiglie residenti in Italia, a partire dalle coabitazioni reali, dalle dotazioni di beni e device e dalle interrelazioni tra i loro membri nel quotidiano. Un formidabile strumento di conoscenza – lontano da stereotipi e luoghi comuni e profondamente diverso dalle letture meramente statistico-giuridiche o teoriche – da cui scaturisce un racconto quasi intimo che rilancia l’immagine di luoghi di vita ricchi di relazioni e oggetti, con persone alle prese con un cambiamento epocale di abitudini individuali e relazionali indotto dalla penetrazione capillare di smartphone e web. Un Rapporto che obbliga a misurarsi con la vita vera delle famiglie residenti in Italia e che permette di capire: 1. chi vive con chi, ossia le convivenze effettive nelle abitazioni individuate grazie al criterio di rilevazione utilizzato nell’indagine (prende in considerazione le persone che convivono per almeno sei mesi e un giorno l’anno); 2. le “cose nelle case”, ossia la fotografia dell’insieme di oggetti contenuti nelle case degli italiani (oltre alla televisione, ancora vera regina della casa, emerge l’ampia e articolata dotazione di device e dispositivi di vario tipo, anche in relazione al grado di connessione con internet); 3. utilizzi e decisioni, ossia come sono utilizzati i device presenti nelle abitazioni, con particolare riferimento alle modalità della fruizione, se individuale o collettiva; e, poi, l’attribuzione del potere decisionale in famiglia sulla destinazione delle risorse familiari tra vari ambiti di consumo. Ampiezza, articolazione e originalità dei risultati permettono di affiancare il Rapporto ai dati e ai lavori sugli stessi temi delle più autorevoli fonti istituzionali, quali Istat e Banca d’Italia.”>

The Auditel SuperPanel™: what it is and how it works

The 16,100 households included in the Auditel SuperPanel™ are extracted through an anonymous and random procedure thanks to the outcome of the Auditel Establishment Survey. Thanks to its composition and weighting system, the Auditel SuperPanel™ correctly represents the wider target population it comes from. Representativeness bases on two factors: a system of enrolment cells matching geographical variables (city size and area), structural characteristics of households (age of head and number of members) and number of TV sets; a double expansion system (one for household data and one for individual data) consisting of a pre-expansion stage for cells and iterative marginal weighting. The Auditel SuperPanel™ distribution on the 103 Italian provinces is proportional to the population, while the panel dispersion covers over 3,500 out of the 8,000 Italian municipalities. The panel households are equipped with a meter, an electronic device which automatically detects the TV channel being watched. The meters measure the TV viewing of the household members and any possible guests minute by minute every day.

Auditel began measuring TV audience in 1984. The sample size was doubled in August 1997 and tripled in June 2017 following increasing audience fragmentation: the Auditel SuperPanel™ is nowadays composed of 16,100 collaborating households. Three main objectives have been reached with our SuperPanel™: 1. increased granularity when measuring small, thematic and local broadcasters; 2. greater stability when measuring bigger broadcasters: the bigger sample size has permitted to reduce the expansion factors by two thirds as well as the so-called confidence interval; 3. lastly, the Auditel SuperPanel™ enables us to take an accurate high-definition snapshot of the Italian society and its socio-economic classes, defined according to dynamic and ordinal logic instead of the mere possession of goods. In conclusion, the Auditel SuperPanel™ represents a forefront at international level and is an absolute point of reference among the Europeran JICs -Joint Industry Committees, the organizations measuring TV audience – not only for its sample size, but also and above all for the quality of its enrolment, management and control protocols. As a matter of fact, the Auditel SuperPanel™ is today the only “engine” in the world that can measure both traditional TV and streaming TV audience, playing the role of the so called ‘Single Source Panel’, that is the unique source which will integrate the two measurement systems – sample and census – in the future. Besides measuring traditional TV with greater accuracy and in finer detail, the Auditel SuperPanel™ also profiles, validates and deduplicates the census results, so that big data can be finely adjusted and a cross-platform and cross-device datum can be obtained avoiding deduplication caused by any possible overlapped viewings: a fundamental datum for reliable audience measurement on all platforms and devices. Nowadays, the Auditel SuperPanel™ is among the world’s biggest TV survey samples (for its sample-population ratio) and the major investors in control activities. It relies on 30,540 meters on as many TV sets and devices, that measure the choices of 41,000 individuals all day long, (sophisticated analysis and control procedures collect and validate the data collected by the meters). A dedicated process examines every single individual and calculates their expansion factor. This system weighs the sample units in order to obtain correct estimates in line with the relevant parameters, so that Auditel can produce a reliable representation of the whole TV consumption choices in Italy.

The sample system: what it is and how it works

The 16,100 households of the so-called Auditel SuperPanel™ sample (representing the entire population residing in Italy) are equipped with an electronic device (the “Meter”) for monitoring their TV consumption on traditional TV, Smart TV, PC, Game Console, etc. 24 hours a day. A complex technological architecture collects, weights and releases the audience data that is published daily just before 10am. Following data processing, the bureaux active in the Italian market allo broadcasters, advertisers, media buyers, etc., to perform sophisticated analyses. In fact, Auditel provides valuable in-depth information on TV viewers (such as gender, age, geographic area, socio-economic class, education) and programme performance (share, minutes viewed, peaks, falls, etc.), that are precious for both publishers and advertisers. This means that every day Auditel produces a currency recognized and used by the TV market to manage business relationships between the producers of TV content and the buyers of TV advertising slots.

The sample which forms the Auditel SuperPanel™ is the sum of two samples: the People Meter Panel (PM) and the Set Meter Panel (SM). Let’s look at the difference.

The People Meter Panel (PM) counts 5,682 households (updated according to the universes valid since August 2018). This number includes oversampling applied to those provinces whose number of households, following proportionality criteria, would have been less than 20 (minimum threshold per province). Additional oversampling has been applied to the regions of Molise and Valle d’Aosta.  The Set Meter Panel (SM) counts 10,294 proportional households. Both the SM and the PM include households without TV and households with at least a non-Italian member. Their TV sets, are called “installed” households. As for sample consistency, the number of collaborating households differs from the number of households whose data are actually collected to produce audience indexes on an average day. The number of households actually producing audience indexes (“households in production” or “validated households”) is lower than the total number of installed households and varies daily. This difference is partly due to households “under repair” having technical problems with their Meters or with the connection to the data collection system. In these cases, the households are temporary excluded from the production cycle while waiting for the maintenance technician to fix the problem. There are also some households whose data are regularly collected but are rejected by the validation system. On average, more than 15,000 households are validated and produced every day.

The Golden Rule: what it is and how it works

Since 16^th December 2018 Auditel has integrated the sample survey system with the census survey system for measuring digital devices. It measures TV viewings from home and out of home on Smart TV, Tablet, PC, Smartphone and Game Console with sophisticated and reliable technology (SDK + Focal Meter) detecting real viewings – stream views – and excluding stream starts and stream requests. A digital tag – or SDK –, installed directly by Auditel on all the players broadcasters provide content from, measures the viewings of any editorial or advertising content played and actually watched by individuals on each device, with no intermediation, second by second.

The Auditel solution doesn’t replace the sample system: it integrates it instead. The next steps are going to allow sample and census data matching, hence the creation of the so-called Total Audience: the aggregation of traditional TV audience with digital device audience thanks to univocal and homogeneous metrics available with both digital and traditional TV standards. The census system detects TV audience on all viewing devices connected through IP protocol (Tablet, Smartphone, Smart TV, PC, Set-Top-Box, Mini-Set-Top-Box, Game Console, etc.). We currently measure web, mobile browser and Smart TV audience. Data from 25^th June 2019 don’t include app viewing information yet, as it is being certified. Six publishers gave their consent right away and are being measured with the census system: DeA, Discovery, La 7, Mediaset, RAI, and Sky, representing 86% of the TV audience. Auditel guarantees free access to the measurement system applying maximum inclusivity and non-discrimination principles. A digital tag, a small software called SDK, has been installed directly inside the players broadcasters provide content from: it measures viewings of any single editorial or advertising content played and actually watched by individuals on each device, with no intermediation, second by second . Auditel is the owner of the digital tag and any rights over installation, development, and operation testing, in compliance with strict certification protocols regulated by the Communications Regulatory Authority (AgCom). We measure both editorial and advertising content. The five editorial content types are:

Live. Real-time viewing of content (or linear viewing, except for buffering).

VOD Long Form. Video On-Demand for deferred viewing of a whole programme after its linear broadcasting.

VOD Short Form. Video On-Demand for viewing clips, derivatives or highlights of a programme (e.g. a football match).

Digital first. Content designed for digital viewing, not yet broadcast on TV, that may be aired at a later time.

Preview. Content available for entire or partial digital viewing before linear broadcasting.

Auditel measures all current digital advertising formats (pre-roll, mid-roll, post-roll) as well as embedded ads within mainly live flow. This complex and sophisticated measurement system is the dynamic source of a library listing all TV contents, including advertising. Every single program and advert carries a unique identification tag (a sort of number plate) for easy identification and tracking in the whole digital ecosystem at anytime.  At present, the metrics we use to measure viewing on digital devices cannot be added together with traditional metrics – they can only be used side by side. These metrics are:

AMR-D (Average Minute Rating-Device). Audience per average digital minute. It’s actually identical to traditional TV AMR, but it quantifies the number of devices (used by at least one viewer) per average minute instead of the number of viewers. Although adding AMR and AMR-D together wouldn’t be correct at present, AMR-D is undoubtedly the closest datum to the “additional share” a programme or channel can get on digital devices.

LS (Legitimate Stream). This metric will probably drive advertising planners’ decisions in the near future. It measures the number of content streams played and viewed on each device for at least 300 milliseconds (the technical threshold ensuring the stream was actually started). It’s calculated for both live and on-demand (VOD) content viewing.

TTS (Total Time Spent). It is the result of the sum of the seconds each device viewed editorial and advertising content of a single channel.

ASD (Average Stream Duration). It calculates the average length of a stream.

Digital device audience data are complex, with endless processing possibilities. This is why the daily data are provided every day at 6pm by the software houses only (Nielsen, Techedge and MCS, that also process sample data), while weekly data are issued for everyone on Tuesdays at 10am.