Background

Although TV attribution is growing in popularity, it is not new. Direct response advertisers have been using simple ‘spot matching’ routines for around 20 years. These spot matching routines typically used a 5-10 minute window or ‘response curve’ which followed a TV spot transmission to match the spike of phone traffic that followed a spot transmission back to that spot. As phone response was usually received on one unique number used for each TV campaign it was a relatively simple task to match that single file of time stamped response data to a spot transmission schedule.

Using attribution to understand how TV drives web traffic

Today, as more and more brands invest in TV advertising to drive web traffic, the focus is on using attribution models to explain how TV spots drive web traffic. However, this is a much more complex area than analysing phone response.

The main challenge is that all brands receive web traffic from a wide variety of sources, 24 hours a day, seven days a week and when paid media advertising is either running or not running. Just a quick look at a Google Analytics report will show you how many sources drive your web traffic:

• Organic search
• Paid search
• Direct visits
• Referrals
• Affiliates
• Display campaigns
• Paid media campaigns
• Revisits

Which traffic do we analyse?

Let’s look at what TV viewers do when they see an ad. They are likely to do one of three things:

1. Enter the brand address directly into their browsers or
2. Click on a paid search (PPC) link or
3. Click on the top organic link.

Reflecting these behaviours, most brands look at:

1. New user traffic through direct browser entries
2. New user traffic through paid search
3. New user traffic through organic search

You will notice that there is a focus on new users. Clearly, new users are of more  interest to brands targeting new customers.

Identifying the baseline web traffic

Identifying the base is complex. This is because a campaign can have a number of baselines depending on the time sample you are looking at. Each hour of the day may have a given level of “natural” traffic. Each day of the week may also have a given level of traffic (this is often the case) and weeks and months may have repeating patterns. Over and above this the brand may have a long-term upward trend in web traffic where each week increases slightly on the previous week. All this means that applying the same baseline to all your analyses will make your results flawed.

The answer to this problem is to use a model which incorporates different baselines based on different times of day, days of week etc.

How does the TV spot matching algorithm work?

Algorithms are mathematical equations that allow a number of variables to be considered simultaneously. So, for TV attribution we need an algorithm that considers the following:

1. The seasonal base
2. The trend
3. The weekly base
4. The day of week base
5. The hour of day base
6. The time of the spot transmission
7. The volume of audience delivered in the spot transmission
8. The time the response is received
9. The way the response distributes over the time period following the spot transmission (the curve)

With this algorithm in place it is possible to calculate a probability that a new web visit that occurred within say 7 minutes of spot transmission was caused by that spot transmission. This process is then repeated across all the spots in the campaign until a probability for all new traffic response to be driven by the TV activity has been calculated.

What type of reporting is available through TV attribution?

Because we have the attributes of the spot (TV station, date, day of week, time of day, type of break, type of creative etc) we can report all these metrics on an aggregated basis. So, for example, we can say Fridays are the most responsive on a % response rate basis, or the most efficient on a £ CPA basis. We can also say which channels and time of day or most responsive.  With this insight we are able to optimise the TV buy to focus budget into the station, days of week and times of day that will deliver the highest ROI.

You can read more about optimising DRTV campaigns here

The post What is TV attribution modelling? first appeared on Marketing IQ.

To illustrate some of the key differences here is a paid media summary in the context of TV:

Objectives:

• Brand advertising tends to seek a change in attitudes towards a brand and deliver uplifts in “lower funnel” sales channels such a display, search and social media
• Direct response advertising tends to seek an immediate behavioural response – the generation of immediate clicks, leads, sales or donations.

Creative strategy:

• Brand advertising tends to position products and services relative to each other in their category and differentiate them using emotional involvement and engagement.
• Direct response tends to persuade consumers to buy immediately using rational messaging.

Here’s a brand advertising TV creative example: Brand advertising building emotional connections – Moneysupermarket

You can see how the essence of the Moneysupermarket ad is entertainment – it uses striking imagery to make an impression on you, build an emotional connection and increase brand trust. The aim is to increase your emotional preference for the brand and reduce your reliance on the functional benefits of the product. That way, when it comes to conversion you will opt to buy from a brand you’ve heard of, feel connected to and trust – even if the pricing or functional benefits are not necessarily the best in the market. In the case of Moneysupemarket, the “do you feel epic?” line invites consumers to be part of a movement.

Here’s a direct response TV (DRTV) advertising example: Direct response advertising is looking for an immediate behavioural response – clicks, quotes, calls, leads or sales

Here you can see how DRTV aims to deliver short-term behavioural change – i.e. web visit response – by covering a lot of selling points in a very short period of time. There is no attempt to gain an emotional connection through entertainment. Quite the opposite – here the intention is to persuade consumers using rational argument.

Ad Timelengths:

• Brand advertising can work on lower timelength edits – typically these are 30 seconds or less – 20s or 10s.
• Direct response advertising tends to require longer timelengths to allow the persuasive arguments to be built and the call to action delivered.

Media Frequency:

• Brand advertising requires both reach and controlled repetition to drive memory. Typically this might be 80% reach at 5-8 OTS  –  that requires between 400 and 640 TVRs.
• Direct response advertising aims to maximise reach at lower levels of frequency so TVR weights can be mush lighter. Given that in the UK, 10 adult TVRs equates to 5m impacts, this weight is adequate to test the responsiveness of an ad.

Media Dayparts and Programme Type:

• Brand advertising requires access to working target audiences which means advertising when they are available to view –  typically this is when they get home from work post 5.30pm – otherwise known as peak. Tends to require high quality programme content environments to maximise chances of engagement with advertising.
• Direct response advertising tends to work best in low interest programme environments and in dayparts where airtime is less demanded and therefore less expensive  – this tends to push DRTV advertising into off peak airtime.

Media Weight:

• Brand advertising tends to require heavier campaign weights. This is because of the requirement to build reach and frequency. There is also strong evidence that share of voice can correlate positively with share of market outcomes
• Direct response aims to maximise reach on the basis that consumers who do not respond on the first or second exposure are unlikely to respond to subsequent exposures in the short-term.

Campaign Evaluation:

• Brand evaluation is based on its objectives – typically these are awareness and consideration shifts and uplift effects on other media channels such as display, search and social.
• Direct response advertising tends to be evaluated based upon immediate response metrics,. clicks, calls, leads, sales, subscriptions and donations
• You can read more about evaluation here

The post How is brand advertising different to direct response advertising? first appeared on Marketing IQ.

Second party data is data you share with a known and named partner. For example, if you are a hotel group you might exchange data with an airline to improve your targeting models;  might add data (sometimes called appending or augmenting) from its airline partner to improve its targeting model. The appended airline data might reveal that a customer always travels business class by air but always books an economy room thus presenting an opportunity for cross-sell. The data added by this cross-party transfer improves the level of insight that can be generated about a given customer and presents commercial opportunities on both sides. This data sharing is enables by the consumer if they tick a data sharing box in a permission request.

Third party data is data that does not belong to you but can be bought or used by you to improve insight or targeting. This is usually sold by third party data suppliers such as Acxiom or Experian. This data has been sourced directly from the consumer and permissioned through an opt-in. Third party data is often used in “matching” projects where a first party database is matched to a third-party database (like the way second party data is used in the scenario above) which can add incremental information to that already held by the first party data owner. So, for example, if you are a retailer of clothing you might want to match your database to a database of clothing purchasing habits to target consumers with products which appear to be relevant to the third consumer base.

Impact of GDPR on first, second and third party data

The General Data Protection Regulations (GDPR) will affect all three types of data and all the companies who are storing and managing that data. Companies holding first party data will need to make sure that their data is properly stored, consented, encrypted and secured in order to meet the regulations. And whilst there is a strong onus on first party data holders to comply with GDPR they, as the data owners, are in a strong position to comply because they are in control of their own data, storage environments and protocols.

The real complications arise when we look at second and third-party data. As first party data is shared with second and third parties, the responsibilities of the first party data owner “stretch” as far as the data goes.  So, if a second or third party commits a breach involving your data, you may still be responsible, at least at joint level with the party you have shared to.

This means you will need to ensure that the way your data is used after being passed to a second or third party remains compliant all down the line. It may not suffice to have your partners sign an agreement saying they will manage the data in line with the requirements of GDPR. If they do not, you may still be liable.

The post What is first, second and third-party data and how is it affected by GDPR? first appeared on Marketing IQ.

Unfortunately, many marketing and media channels are planned, negotiated, delivered and evaluated in silos. This means it can be difficult to get a set of comparative response rates which allow you to forecast how well any one channel may work for your business or brand. If you can’t compere them side by side it’s difficult to optimise budget distribution – particularly for customer acquisition activity.

Guide to response rates by media communication channels

With over twenty years’ experience of planning, managing and evaluating campaigns across practically all mainstream media channels, I thought it would be useful to share the metrics that I use as standard response metrics. These are given as percentage response rates of the audience seeing the ad.

Note: These are the response rates I would expect to see based on my experience. They should be used  as a guide and are not a guarantee. They are subject to the caveats listed below.

The caveats

1. Response rates are driven by a number of factors including the product, offer, the creative treatment and the audience selection (media). Ideally, you should work to the highest possible standard in each of these four areas. Compromise on any of these factors will reduce response rates.
2. Most channels have sub-sets of response rates depending on how the channel is being used. For example, TV ads can be “brand awareness” ads, “brand response” ads or “direct response ads”. Each of these have different levels of responsiveness. Brand awareness ads which are designed to change attitudes rather than short term behaviour will not deliver a high response rate.
3. You must factor in the cost of media on a per audience basis. A favourite mistake of response rate observers is to look at response rates without factoring in channel costs. Here’s an example; the response rate from DRTV is about 100 times lower than the response rate from DM, but remember, DM costs around 100 times more per person than TV. In reality, both channels may produce a similar cost per response. That’s why it’s important to look at both factors when analysing and forecasting responses.
4. Response rates aren’t everything; what generates revenue is sales so you need to factor in a conversion rate from response to sale.  As a general rule, personal channels like DM tend to convert at a higher rate than broadcast or online display. You can have a channel with a low response rate and high conversion rate performing as well in cost per sale terms as a channel with a high response rate and a low conversion rate.
5. Marketing activity is subject to diminishing returns; response rates will fall as budgets increase.

For more information on our High Performance Media Planning course please follow this link

The post Advertising Response Rates by Channel first appeared on Marketing IQ.

Warm Direct Mail – mailings to your active customer file: In our experience, warm direct mail, i.e. DM sent to your customer file should deliver a response rate of between 1% and 5%. The average figure is around 3.5%.

Cold Direct Mail – DM send to prospects via a “cold” list: Response rates here are lower as the consumers you are mailing are less familiar with you and your brand. Typically 0.5% to 1.5%.

The DMA in the UK cites a response rate of 4% and claims that overall 7% of recipients will take some kind of action as a result of receiving direct mail.

The DMA in the US has produced a lot of information in its 2015 Response Rate Report and cites response rates of 3.7% for a house list and 1% for a cold prospect list.

The post Direct Mail Response Rates first appeared on Marketing IQ.

The most common forms of multivariate analysis in marketing are cluster analysis and hierarchical analysis. Cluster analysis uses statistical techniques to allocate customers into segments based on how similar, or dissimilar, they are to each other. So for example, if you had 10,000 customers and you were clustering by income and home ownership, you would be able to define groups of customers with similar levels of income and home ownership status, or those with high income and low home ownership status, or those with low income and high home ownership status. The number of clusters generated depends on how you set up your cluster analysis and of course, what patterns actually lie within your data. You can set up your analysis to produce either a large or small number of clusters, but most marketers can’t practically service more than about fifteen clusters.

Hierarchical analysis breaks customers down into sub-sets of the whole customer base. Results of hierarchical analysis are often shown as dendrograms or tree diagrams. In a tree diagram, all customers belong to the ‘root’ and segments of the customer base are called ‘nodes’, nodes are connected to the tree by ‘branches’.  So for example, all customers can be divided into males and females. Then the males and the females can be divided by age, and then by income and then by spend. You are then able to see what proportion of the whole base is composed of customers with certain characteristics.  Here are some examples of customer segments defined using hierarchical analysis:

1. Spend more than £250 per year and are aged 18-34 and female and do not have children
2. Spend more than £500 per year and are aged 25-44 and male and do not have children and earn between £20,000 and £30,000 and have a mortgage
3. Spend more than £1000 per year and are aged 35-54 and have children and have a mortgage and live in the South East

Whichever technique you use, it is likely that you will see a small number of segments account for disproportionally large amounts of sales revenue or sales potential. When you have identified these segments you can leverage what you know to develop tailored offers, messages and targeting. Over and above this you can identify customers who have the characteristics of high performance segment membership, but are not spending at the rate they could be. You can use this information to target your marketing messages to the sales prospects with the highest untapped potential.

The post How is multivariate data analysis used in marketing? first appeared on Marketing IQ.

Here’s an example: Let say you sell 10 products. It may be the case that all purchasers of product 8 are: 1) in a certain geodemographic group, 2) married with more than one child and 3) own more than one car. All these factors can be analysed and combined to predict the likelihood of any consumer in your database buying product 8. Usually this combined measure is referred to as a ’score’ i.e. a figure which represents the presence or combination of certain variables in the consumer record. Once you have developed your scoring model you can rank all customers by their score. When you’ve stripped out those who have already bought product 8, you are left with a set of high potential prospects.

Predictive modelling can also be undertaken based on transactional information about past purchases. Going back to the 10 products, it may be the case that 80% of people who buy product 7 have previously bought products 2, 5 and 6 and in that order. So we can say that people who have bought products 2, 5 and 6 (in that order) but who have not yet purchased product 7, are much more likely to buy product 7 than everyone in your database. Again a score is attached to these behaviours and that score can be used to rank your prospects in terms of untapped sales potential.

Of course as well as predicting purchase behaviour, these techniques can be used to predict risk. In credit assessment for example, it may be the case that those customers who have certain demographic characteristics combined with a certain type of past purchase behaviour are highly likely to default on a credit agreement. This is sometimes referred to as credit scoring. If you are rejected for credit at a bank or in a shop it will be because your data has been analysed and your credit risk score is deemed too high or low to meet the criteria of the lender.

These predictions can help you target your communications very efficiently and also help you control commercial risk in customer behaviour. What’s interesting about these techniques is that they help both the marketing department and the finance department. Marketing delivers customers who are both highly likely to convert to sales or high lifetime value whilst at the same time, producing customers who are less likely to cause problems for the finance department. Overall, this means that the resources of the business are being better utilised.

The post What is predictive modelling in marketing? first appeared on Marketing IQ.

At the most basic level your database should contain a name and address for each record. The name and address can give you valuable information. The postcode in the address opens up the potential for geodemographic analysis using tools like ACORN or MOSAIC. These tools work by grouping consumers into clusters of similar people based on the types of neighbourhoods they live in. The principle behind these systems is simple; birds of a feather flock together. The owners of these segmentation systems undertake research into the clusters they have developed. For example, Cluster 1 may contain people who are known to be affluent pre-retirement couples with children who have left home. Research may show that these people are three times more likely to drive a certain car, purchase certain electrical products or take holidays to certain destinations. So from just the address record you can build a much wider picture of the record in question.

But the full name and address have even more potential. They can be used to match your customer file with an external data file containing more information about the same person. This data can come from many sources, but more often it comes from lifestyle surveys. If a customer in your database has completed a lifestyle survey then you can buy supplementary information to significantly expand what you know about that person.Here’s an example. You may only know the name, address and age of a customer. But if that record can be matched with a respondent to a lifestyle survey then you can see the answers to tens or even hundreds of other purchase preference questions that person has shared. For example, you may be able to see what type of car they own, when it was bought, when they intend to replace it. They may even tell you what type of car they are considering next.

If you have transactional data then you are able to undertake an analysis of the types of products and services bought by the customer. From this data you would be able to say that a customer owns products X, Y and Z and you will probably know when they bought those products. You will be able to see how the often products are purchased and the preferred means of payment. If there is cyclical behaviour in the purchase pattern you may be able to predict when this customer is likely to purchase those products again.

With these high levels of customer understanding you are able to take a lot of the guesswork out of marketing. You can be much more focussed in terms of selling specific products to specific individuals. As a result you response, conversion and customer value rates are likely to improve significantly.

The post What can a database record tell us about customers? first appeared on Marketing IQ.

• Exactly who buys a given product or service
• Detailed information about the characteristics of those buyers
• Which other products and services they buy
• Which product and service offers they find most attractive
• Which buyers buy more of certain types of products
• How you can find more buyers with the same characteristics

These data analysis techniques can be applied to all types of customer data – whether it’s for a retail business, an online business or a call centre based business. Insight from data analysis can be applied across a wide spectrum; from adding inspiration to a creative brief through to changing a company’s entire business strategy.

You may think the claim that data analysis can change the destiny of a business is rather grandiose. But I can can think of two examples of breakthrough data insight from the same category that ended up contributing millions in additional brand revenues.

Sainsbury’s  – Sainsbury’s agency AMV were tasked with increasing the then ailing retailer’s sales by £2.5bn over a three year period. A seemingly impossible challenge until viewed as a data question. The AMV team calculated that £2.5bn equated to £833m per year which in turn equated to £16m per week.  It still looked like a big number until the AMV team considered that Sainsbury’s handled 14m customer transactions per week.  Then the target equated to just £1.14 per transaction. The brief to increase sales by £833m per week could be redefined as increasing each existing transaction by just £1.14. Now the target not only looked attainable, but this data insight led to the idea that lots of small changes could make a big difference.  From this insight came the campaign idea that consumers should “Try something new today”. By asking customers to ‘try something new’ they were able to persuade customers to spend at extra £1.14 every time they shopped.

Tesco – The Tesco Clubcard is now legendary as both a customer loyalty card and a source of information about customers.  Up until the introduction of the loyalty card, many retailers didn’t know who their customers were. And if they didn’t know who they were it was difficult for them to gather the data that allowed them to understand individual customers better. With the Club Card this all changed. Tesco were able to develop individual data driven relationships with their customers.  They were able to understand customer needs better and in doing so they gained competitive advantage over their rivals.

The post Marketing data analysis gets you closer to customers first appeared on Marketing IQ.