practice papers beyond open rate: why it is time for e ... · alchemy worx, 4th floor, 1 new oxford...

INTRODUCTION

E-mail is not digital direct mail

The e-mail marketing landscape hasnever been healthier. In the last threeyears, the number of e-mails sent bybrands has nearly doubled with nodiscernible change to open or click rates(Figure 1). Information held by the UKDirect Marketing Association indicatesthat consumers are increasingly choosingto receive a greater number of e-mailsfrom brands they trust. In the recent USpresidential election campaign, Obama’sdigital fundraising was ten times higher

than Romney’s because he sentsignificantly more e-mail.1 But despiteoverwhelming evidence that consumersare willing to receive more e-mail fromtrusted sources, marketers are still tryingto get subscribers to open, click and buymore often, by sending fewer e-mails.E-mail inherited much of its strategy andthought leadership from direct mail,where the biggest and defining challengeis the high cost of sending messages. Thehigh channel costs of traditional directmail led to the development of the RFMmodel (recency — frequency —monetary value), which focuses on

� Henry Stewart Publications 2050-0076 (2013) Vol. 1, 1 39–55 Journal of Digital & Social Media Marketing 39

Practice papers

Beyond open rate: Why it is timefor e-mail marketers to thinkreach, frequency, impactReceived: 14th February, 2013

Dela Quistis founder and CEO of Alchemy Worx, the leading digital marketing agency with 100 per cent focus one-mail. Dela is a popular international speaker on e-mail, and his thought-provoking views and livelystyle regularly receive rave reviews. Dela has served many years as a member of the UK DirectMarketing Association’s E-mail Marketing Council and chair of the Email Marketing CouncilBenchmarking Hub. He has also served on the executive management group of the Internet AdvertisingBureau and on the steering committee of the Future of European Advertising Stakeholders.

Alchemy Worx, 4th Floor, 1 New Oxford Street, London WC1A 1LW, UKTel: +44 (0)20 7025 2100; E-mail: [email protected]

Abstract E-mail marketing is still dominated by a direct mail world view, with a focus onrecency, frequency and monetary value (RFM) and a resulting fixation on relevance, timingand tighter targeting. As a result, e-mail marketers are trying to get subscribers to open,click and buy more often by sending fewer messages. Current performance metrics areboth limiting and misleading. Thinking of e-mail as fundamentally a broadcast advertisingchannel shifts the focus onto reach, frequency and impact (RFI), unlocking the brandpotential of e-mail.

KEYWORDS: e-mail marketing, reach, frequency, brand

Quist:JSC page.qxd 10/05/2013 10:52 Page 39

relevance, timing and targeting tomaximise return on investment (ROI).By continuing to adhere to this approach,e-mail marketers are not leveraging thetrue power of e-mail — its low cost —to test consumer acceptance forfrequency and unlock the brand potentialof e-mail.

Unlocking the brand potential of e-mailIn every other marketing medium,advertisers try to push the boundaries offrequency to see how much messaging anaudience will accept. Imagine if ad timeon television was free. Would advertisersreduce, maintain or increase their numberof ads? They would take as much space asthey were offered because they use thebroadcast RFI model (reach — frequency— impact). Importantly, it is this model,not RFM, which brand marketers haveadopted because it focuses on telling asmany people about a brand as many timesas they will tolerate. With the lowest costper message of all marketing channels,e-mail would appear to be the perfectcandidate for the RFI model. However, it

is hugely underexploited by marketersbecause of the perception that inboxes arealready overcrowded.

The (un)crowded inboxAny discussion about e-mail marketingwill inevitably mention over-mailing orspam. Genuine spam is defined asunsolicited e-mail. It is illegal and verylittle of it makes it into inboxes these days,such is the quality of the filters thatinternet service providers have in place.Nearly everything else is legitimate,permission-based e-mail from trustedbrands, friends and colleagues. It isimportant to distinguish commerciale-mail from personal e-mail becauseconsumers do this very effectively forthemselves. In Merkle’s ‘View from thedigital inbox’ report,2 when people wereasked how they would like to becontacted for personal communication,approximately 35 per cent said e-mail,with phone and text message provingmore popular (Figure 3). When asked howthey would like to be contacted forcommercial communication, e-mail is the

Time for e-mail marketers to think reach, frequency, impact

40 Journal of Digital & Social Media Marketing Vol. 1, 1 39–55 � Henry Stewart Publications 2050-0076 (2013)

Figure 1: UK DMA National E-mail Benchmarking Report 2011


overwhelming choice, with roughly 75 percent of people choosing this as theirpreferred option (Figure 2). Most peoplewould rather be sold to via e-mail than byphone, text or social media.

In reality, most customers do notexcessively love or hate e-mails frombrands they trust. They like the brandenough to tolerate the e-mails until suchtime as they require the service or productbeing offered by that brand. We call thisunemotionally subscribed. Such apparentindifference should not be mistaken fordislike or disengagement. E-mail is asufficiently mature marketing channel forconsumers to understand that they alone

choose what arrives in their inbox. If theyhave not unsubscribed, it means they wantto receive e-mails from that brand, whetherthey engage with every e-mail or none atall. Interestingly, people’s view onfrequency and volume is directlyconnected to their opinion of the brand.Brand advocates will happily receive twoor three e-mails a day, while brandopponents might see two e-mails a week asover-mailing. E-mail is the tail not the dog— annoy people from a brand perspectiveand they will start shouting ‘spam’.

Finally, the most sophisticated filter fore-mail is the consumer themselves. Atypical consumer will choose to receive


Quist

Figure 2: Preferred method of commercial communication by age

Source: Connelly, L. (2011) ‘View from the digital inbox 2011’, available at:http://www.merkleinc.com/thought-leadership/white-papers/2011/view-digital-inbox-2011 (accessed 18th March,2013)




e-mails from one or two competingbrands only, making the inbox a relativelycompetition-free zone. Take insurance asan example. Searching on Google wouldreturn hundreds of insurance offers, whilea search of a consumer’s inbox wouldlikely find just one or two — those towhich they are subscribed. One or twoe-mails from an insurance company aweek does not contribute to anovercrowded inbox. According to theDirect Marketing Association’s e-mailtracking report for 2012, 57 per cent ofUK consumers subscribe to ten brands orfewer. If each of these brands sends three

e-mails a week, a high frequency byconventional e-mail wisdom, that wouldmean that over half of UK e-mailconsumers were receiving a maximum of30 commercial e-mails a week, most a lotless. Inboxes are only crowded frommarketers’ perspective.

BEYOND OPENS AND CLICKS

Focus on customers not campaignsThere is a fixation in the e-mail industrywith rates and, more particularly, open andclick rates. These are widely used as the

Figure 3: Preferred method of personal communication by age

Source: Connelly, L. (2011) ‘View from the digital inbox 2011’, available at:http://www.merkleinc.com/thought-leadership/white-papers/2011/view-digital-inbox-2011 (accessed 18th March,2013)

Quist:JSC page.qxd 10/05/2013 10:52 2

definitive performance indicators but canbe very misleading. In Figure 4, there is aclear spike of about 50 per cent in openand click rates for weeks 12 and 13. Takenin isolation, this would appear to be asignificant success and it would seemlogical to replicate the creative, product,offer or targeting. As shown in Figure 5,however, it was a highly targeted, tinyvolume send that produced the outlier.Unsurprisingly for such a tiny sendvolume, revenue was also low. As the keyobjective for almost every business isrevenue, this is what should drivemeasurement and strategy.

Open and click rates are also responsiblefor one of the most divisive disconnects inthe e-mail industry, namely the use ofcampaign metrics to measure customer

behaviour or engagement. Behaviour inmarketing is defined as the aggregation ofspecific responses by customers to astimulus, in this case an e-mail campaign.Campaign metrics measure these specificresponses, such as open and click, but donot offer any insight into the relationshipbetween them. For example, if an e-mailnewsletter has an average open rate of 21per cent after five sends, does that meanthat only 21 per cent of the database hasengaged over that time? It is far morelikely that there is a group of subscribersthat open every e-mail and a group thatnever open, with everyone else sittingsomewhere in between. This is a valuablepiece of information about customerbehaviour that cannot be gleaned fromconventional metrics because they do not


Quist

Figure 4: Open and click reach vs open and click rate


examine the relationship between theseresults over time.

What adds to this campaignmeasurement fallacy is that mostcompanies do actually requireperformance to be charted over time andfor many brands, excluding fast-movingconsumer goods, customer purchase cyclesare longer than one month. Thus, there is

a need for a metric that keeps track ofcustomer behaviour over time. This metricis one of the cornerstones of the broadcastRFI model, and it is called reach.

Reach: Measuring customer behaviourReach is a customer behaviour metric thatmeasures the proportion of a list that has



Figure 6: Open, click and purchase reach vs cumulative revenue for UK travel company

Figure 5: Open reach vs frequency


engaged at least once over a period oftime. In e-mail terms, that interactioncould be opens, clicks, conversions orpurchase, but for simplicity’s sake thispaper will consider only opens. Openreach is calculated by counting thenumber of subscribers on a list that haveopened at least once, and dividing that bythe number of subscribers that havereceived at least one e-mail over a givenperiod, thus:

open reach=

The time period could be a quarter or ayear, depending on send frequency andspecific goals, but to achieve significantresults the chosen period must contain atleast 12 interactions.

First-time opens drive resultsFor open reach to have any practicalapplication in e-mail marketing, it mustcorrelate to the most important businessmetrics of conversion and revenue. E-mailagency Alchemy Worx has been measuringopen reach across a range of clients and

industries for seven years and time andagain it has noted such a correlation. Thisis perfectly illustrated in Figure 6. The firstarrow highlights a period in which totalrevenue, the grey line, was static and this ismirrored closely by open reach, in lightgrey. The second arrow highlights anincrease in total revenue that correlates toan increase in open reach. The bulk oftotal revenue is generated when customerswho have not opened over a given period,open an e-mail for the first time. Thus, astrategy that focuses on optimising openreach instead of open rate is far morelikely to improve revenue.

The reason for the correlation betweenrevenue and open reach is at the heart ofthe RFI model — a very large group witha low response rate will generate morerevenue than a very small group with ahigh response rate. Figure 7 details theannual revenue generated by openfrequency for a UK travel company. It isinteresting to note that the largest groupfor revenue generation is ‘0 messages’, butthe power of unopened e-mail will bediscussed in the branding section a littlelater. Moving up the open frequencychain, the percentage of total revenue


Quist

de-duped list of all e-mailaddresses that opened an e-mail

de-duped list of e-mail addressesthat received an e-mail

Figure 7: Annual revenue generated by open frequency for a UK travel company




generated by the higher responding groupsdrops in line with the diminishing size ofthe group. Although it is easy to focus onthe groups with the best response ratesbecause the open and clicks look sopositive, it is the much larger groups thatgive the biggest increases in total revenuewhen they open. In the RFI model, theshort-term goal of increasing open andclicks is replaced by a longer-term strategy— converting non-openers into first-timeopeners, and then those first-time openersinto higher-responding multiple openers.

Target engagementThe traditional direct mail model usestighter targeting to reduce costs andincrease ROI through relevance but itsapplication in e-mail can becounterproductive. First, the ability tosegment and target effectively is entirelydependent on the quality of data. Evensimple data points such as age, gender andname can be difficult to collect accuratelyfor an entire database and it is gettingharder, too. According to the UK DirectMarketing Association’s most recentnational client e-mail report, the numberof UK client-side marketers citing data asa barrier to achieving their goals rose bynearly one-quarter from 2011 to 2012.Secondly, it is very hard to predict withaccuracy when and why a customer willbe ready to do something new with abrand. Imagine a buyer of fine wine ordersa few cheap cases from a mail-ordercompany every year for the office party.That transaction leads to a data profile thatrecommends targeting him with cheapwine offers. In fact, the fine wine buyer’spersonal tastes are very expensive, butbecause his purchase history excludes himfrom fine wine offers, he buys hisexpensive wine elsewhere. Both are happybecause they continue to gain somethingfrom the relationship but both are alsooblivious to how much more beneficialthe relationship could be.

A more effective and measurable area ofsegmentation is customer engagement withthe communications themselves. Trackingopen reach makes this data available forevery single customer. In terms of ane-mail strategy using the RFI model, thismeans understanding how first-timeopeners and multiple openers respond toe-mails in comparison to each other. Forexample, a free delivery offer is sent to anentire list. Open rates and revenue for thatindividual campaign are both aboveaverage; then a 20 per cent discount offer issent and both metrics are again aboveaverage. Using conventional metrics, bothof these campaigns would appear to besuccessful, so the logical conclusion wouldbe to keep sending these two differenttypes of offer to the entire list.

However, analysis of open reach forthese offers shows that free delivery hadvery little effect on generating first-timeopens, whereas the discount offer hadsignificantly more. A free delivery offer hasless interest to someone who is not buyingregularly but is more likely to appeal tosomeone who is. Likewise, a big discountis far more likely to engage someone whohas been inactive for a while. Analysis ofopen reach can identify what triggersthese different groups to respond so thatthey can be targeted with an appropriateoffer and message. Customers areconstantly indicating how they want toengage with brands and open reach allowsmarketers to listen, learn and lift response.

Reactivate with every e-mail sentThe use of open reach as a keyperformance metric for e-mail identifiesthree key groups — non-openers,first-time openers and multiple openers. Asdiscussed, the first and most important stepof the strategy is converting previousnon-openers into first-time openers. Butlook at this another way and everynon-opener is actually an opted-inprospect in need of reactivation (see the


pale grey section in Figure 8). One-offreactivation campaigns are often ineffectivebecause they target a snapshot of adatabase. A small number of customers willengage but that is often their onlyinteraction until the next time they appearin the snapshot, and so on. Rather thanfocusing one campaign every six or 12months on re-engaging inactives with alow-margin offer, every e-mail should beseen as a re-engagement opportunity.Insights gained from analysing whatmotivates non-openers to becomefirst-time openers will inform thiscontinuous reactivation strategy. Thereshould be no need to send ‘we miss youe-mails’ when open reach analysis suggeststhat a 20 per cent discount offer hassignificantly increased first-time openers.This is a measurable trigger for this groupand can be used as part of a strategy thatreactivates with every single e-mail sent.

The importance of list sizeWhen there is a significant increase inopen reach, typically it is for one of two

reasons — new people were added to thelist or there was an increase in mailingfrequency. Analysis of open reach showsthat new subscribers to a list are the mostlikely to become first-time openers, so oneof the best ways to increase revenue is togrow the list. This is why in the RFImodel, database size is the single mostimportant factor for increasing revenue. Itcan be a large investment of time andmoney, but the impact of new subscriberson open reach, and subsequently revenue,is significant. Very few marketers ascribeenough importance to increasing list sizewhen it comes to e-mail because the focusof the RFM model is reduction.

Frequency increases engagement:Case studyThe second most important factor in theRFI model is mailing frequency.Periodically increasing send frequency isthe most cost-effective way of optimisingopen reach and the single most effectiveway of increasing revenue in any e-mailprogramme, regardless of strategy or model.


Quist

Figure 8: Example reach chart categorising all mailed subscribers


As such, this part of the RFI e-mail strategyis worthy of more in-depth analysis.

The following is a case studyundertaken by Alchemy Worx to test theimpact of frequency on engagement. Theprevailing wisdom in e-mail marketing isthat increasing frequency decreasesengagement. This study of a UK travelcompany’s monthly newsletter shows theopposite to be true. The database wasequally split into four groups, with groups1 and 2 receiving the full newsletter plus are-send for group 2, and groups 3 and 4receiving a bi-monthly e-shot that splitthe content in two plus a re-send forgroup 4 (Table 1). The test was run overthree months to make sure the content ofthe e-mails did not skew results. For thepurpose of this discussion, the two groupswith the biggest variation in frequencywill be considered, namely, groups 1 and4. Group 1 received three e-mails perquarter and group 4 received 12 e-mailsper quarter, four times the send frequency.

The first thing to note about the results(Table 2) is that the average open rate for

the higher-frequency send is significantlylower than that of the lower-frequencysend — a difference of 7 per cent in thecase of groups 1 and 4. There is also asimilar result for click rate. It is at thispoint that open and click rate would beused within the RFM model to concludethat customers are reacting negativelywhen sent more e-mail. This againhighlights how misleading open rates canbe because the figures show the totalnumber of opens for group 4 to be twiceas many as group 1 — 20,944 vs 9,386(Figure 9). Again, it is a very similar resultfor total clicks — 1,870 for group 1 vs3,498 for group 4.

A look at the wider picture confirmsthe same trend: 39 per cent of group 4opened at least once in the three-monthperiod compared with 27 per cent ingroup 1, and 24 per cent of group 4opened more than once vs 10 per cent ofgroup 1 (Figure 10). The one negativeoutcome for group 4 was a churn rate 2per cent higher than group 1. In the RFMmodel, where increasing database size is



Table 1: Frequency test groups

Group Mailable records Mailings

1: Monthly newsletter 19,678 32: Monthly + resend 19,641 63: Fortnightly e-shot 19,646 64: Fortnightly + resends 19,747 12

Figure 9: Open and click volume by test group


not a priority, this could be viewed as areason not to adopt a high-frequencystrategy. However, increasing database sizeis a key component of the RFI model, sothe higher percentage of churn will beoffset by the increased acquisition of newsubscribers.

These results show that it is frequencythat drives engagement, not the other wayround as is currently adhered to in theRFM model. Customers cannot engagewith an e-mail they do not receive. If theword ‘e-mail’ was replaced with‘opportunities to interact’, then thestrategy of ‘an extra e-mail send to 1million customers’ becomes the strategy of‘let’s send another 1 million opportunities

to interact’. This simple re-phrasinghighlights the biggest obstacle to the RFImodel — it is the perception by marketersthat sending more e-mail disengagescustomers and not the fact itself.

Frequency needs to be proportionateto valueIt stands to reason that there will always bea limit to the number of e-mails a singlebrand can send within a given timeperiod. It is not just about how manye-mails you send, it is about the relativevalue of each e-mail. Frequency needs tobe proportionate to value. A daily e-mail ismore demanding on the recipient than aweekly e-mail, which in turn is more


Quist

Table 2: Frequency test results

Sent Open Click CTO Unsub Complaint Group Mailings volume Opens rate (%) Clicks rate (%) (%) rate (%) rate (%)

1: Monthly newsletter 3 59,034 9,836 17.01 1,870 3.23 19.01 0.54 0.091: Totals 3 59,034 9,836 17.01 1,870 3.23 19.01 0.54 0.09

2: Monthly newsletter 3 57,028 8,294 14.82 1,453 2.60 17.52 0.39 0.122: Resend 3 47,951 2,998 6.29 508 1.07 16.94 0.38 0.112: Totals 6 104,979 11,292 10.89 1,961 1.89 17.37 0.38 0.12

3: 1st bi-weekly e-shot 3 56,630 7,951 14.27 1,251 2.24 15.73 0.44 0.093: 2nd bi-weekly e-shot 3 56,191 7,751 14.03 1,219 2.21 15.73 0.35 0.113: Totals 6 112,821 15,702 14.15 2,470 2.23 15.73 0.40 0.10

4: 1st bi-weekly e-shot 3 56,696 7,754 13.89 1,134 2.03 14.62 0.40 0.114: Resend 3 48,313 3,311 6.87 900 1.87 27.18 0.36 0.114: 2nd bi-weekly e-shot 3 56,109 7,192 13.02 982 1.78 13.65 0.30 0.084: Resend 3 48,316 2,687 5.58 482 1.00 17.94 0.28 0.094: Totals 12 209,434 20,944 10.10 3,498 1.69 16.70 0.34 0.10

Table 3: Aggregated frequency test results

Revenueper

Open Click Purchase contact ListMailable Sent reach Opened reach Clicked reach Purchased mailed churn

Group records Mailings volume (%) >1 (%) (%) >1 (%) (%) >1 (%) AOV (£) (£) (%)

1: Monthly 19,678 3 59,034 26.99 9.67 6.88 0.79 1.39 0.23 179.58 5.57 1.58newsletter

2: Monthly 19,641 6 104,979 32.16 16.71 8.57 1.35 2.01 0.31 247.32 6.14 2.02+ resend

3: Fortnightly 19,646 6 112,821 32.25 18.88 9.66 2.31 2.10 0.34 225.78 5.96 2.28e-shot

4: Fortnightly 19,747 12 209,434 38.97 24.39 13.36 3.37 2.11 0.34 243.32 6.17 3.52+ resends




Figure 10: Open and click frequency breakdown by test group


demanding than a monthly e-mail. Asfrequency is limited by value, finding theacceptance level through testing andphasing needs to go hand in hand withunderstanding the value of each e-mail tothe consumer. Thought of in this way,e-mail becomes about influencingbehaviour in a brand way, rather thanabout reacting to behaviour in a directmail way. In other channels this approachis simply called brand advertising.

E-MAIL AS BRAND ADVERTISING

The lowest-cost ad channel availableA typical major UK retail brand is likelyto have an e-mail list of between 250,000and 1 million contacts. To put that inperspective, very few national newspapershave a circulation of more than 1 millionnowadays, and even those numbers comewith significant wastage. This means up to1 million customers who have expressedan interest in a brand. Access to that kindof relevant audience is not available in anyother marketing or advertising channel inthe UK, save for prime-time television adslots that are prohibitively expensive, andthe increasingly difficult to monetise socialmedia. Now compare the cost of regularlye-mailing 1 million people who have saidthey are interested in hearing more abouta brand’s products and services, with thecost of booking a slot in a leadingprime-time soap opera every week.

So the broadcast element of e-mail isvery real, and it is arguably the closestthing in numerical terms to the power thatbroadcast television once wielded beforethe signal went digital and the marketfragmented. With numbers like these, thegoal of every major brand should be tobuild and maintain a list of people whohave opted in to receive their e-mailsbecause every send is an opportunity toadvertise their products, services and brandvalues at the lowest cost available.

The nudge effect

The evidence that e-mail can work asbrand advertising is perhaps most notablein the power of the unopened e-mail. Asdiscussed, the RFM model would viewany unopened e-mails as a failure becausethe send of that e-mail has not beendirectly monetised. Looked at from theRFI viewpoint, however, the results arevery different. Figure 11 shows acompany’s sales performance for afortnight across all channels. The verticalmid-grey area indicates the period overwhich the e-mail was sent, the light greyhorizontal line shows sales from peoplewho did not receive the e-mail and thedark grey line shows people who got thee-mail but, and this is the key, did notopen it. Despite not opening the e-mail,they still went on to purchase via anotherchannel. This is called the nudge effect andit is essentially a way of influencingpeople’s behaviour without a directresponse — or, in other words, brandadvertising.

Just because a user is not opening ane-mail, it does not mean that they are notregistering with a brand. The simplepresence of e-mail in a user’s inbox createsthis powerful nudge effect — a subtlestimulus that brings three distinctadvantages:

• regularly seeing a brand in their inboxmay prompt a subscriber to respond viaanother channel;

• an engaging subject line reminds thesubscriber of any current offers acrosschannels, as well as reinforcing brandvalues and key benefits; and

• regular e-mails keep a brand in thefront of a subscriber’s mind, so that evenif they do not transact today, they maydo so in the future.

What is more, this effect is not justnoticeable in isolation. Figure 12 showsthe revenue across all channels, both


Quist


online and offline for customers who didnot open any e-mails. The revenue ishighest in absolute terms for the segmentthat received seven e-mails.

The nudge effect derives itseffectiveness from ‘e-mail triage’. This isthe act of inbox-sifting that all inboxowners go through to ascertain theimportance of each e-mail and therelevant action. Which messages need tobe open and read at once? Whichmessages can be saved for reading later?Which ones need to be kept? Which onescan be deleted? During this process,subscribers must use their highestcognitive function to make these instantdecisions. Sending an e-mail featuring abrand name in the ‘from field’ alongside asubject line that communicates contentand strong brand benefit can facilitateretention of the message and influence a

purchase decision — even when therecipient fails to open the e-mail. It is asubtle yet powerful opportunity thatmakes an unopened e-mail a source ofpotential revenue across all channels.

E-mail drives sales in other channelsIt is not just unopened e-mail that has abranding effect; e-mail as whole, opened orunopened, drives sales and activity in otherchannels. In Figure 13, the revenue from allchannels excluding e-mail was measuredusing Omniture tracking. The light greybars show the average daily revenue permonth on days for which e-mail was sent,whereas the dark grey bars show theaverage daily revenue for days on which noe-mail was sent. Those days on whiche-mail was sent deliver higher revenuethrough non-e-mail channels. As e-mailrevenue has been excluded, it is clear that



Figure 11: Non-email revenue from customers mailed but that did not open vs not mailed


revenue not attributed to e-mail is present.Figures 14 and 15 split the same data intodifferent channels. In each case, the darkerblock represents the revenue for the groupthat continued to receive e-mail, and in

each case there is an increase across allchannels when e-mail was sent. Therevenue generated directly by e-mail isdirect response in action, while the revenuegenerated not directly attributed to e-mail


Quist

Figure 12: Revenue from all channels from subscribers that did not open, by number of emails received

Figure 13: Non-email revenue by month with and without e-mail


is branding at work. The beauty of e-mailis it can be very effective at both brandingand direct mail, and it does so while beingthe cheapest marketing channel everdevised by marketers.

CONCLUSIONSWHAT? E-mail marketing’s inheritedreliance on RFM from direct mail isoutdated and hindering its potentialimpact as a brand marketing channel, aswell as a direct response channel. Much ofthis stems from historical industryprejudice and superstition, and is notborne out when looking at the numbersbehind what is actually happening whene-mails are sent to consumers.

HOW? A shift in focus to the broadcastmarketing model of RFI unlocks e-mailbrand potential by leveraging its low costto focus on reach, frequency and impact.

WHAT? There needs to be a sea changein how the success and failure of e-mail ismeasured. Campaign-based metrics such asopen and click rates are limiting in theirscope and ignore consumer behaviourover time.

HOW? By looking at reach, e-mailmarketers can learn what is working witheach campaign and use that informationto improve revenue and engagement, aswell as reactivating the inactive sections oftheir list with every e-mail they send.

WHAT? E-mail marketers should betesting the acceptance levels of frequency



Figure 15: First touch average daily sales by channel (October 2012)

Figure 14: Last touch average daily sales by channel (October 2012)


because the perception that consumersreceive too much e-mail is not supportedby the numbers.

HOW? Analysis of reach highlights thatfrequency drives engagement in e-mail.Sending another e-mail is still the mosteffective way to increase engagement andrevenue.

WHAT? E-mail is a powerful brandchannel that increases revenue across allother channels regardless of opens, clicksand conversions.

HOW? E-mail marketers should beworking closely with their other channelsto optimise for this behaviour and put inplace a brand strategy for subject lines andcontent.

FINAL THOUGHTFor those that remain sceptical that theRFI model has any real-world applicationfor e-mail, consider these two e-mailstrategies: Campaign R sent an average of20 million e-mails a day with an averageread rate of 8 per cent, while Campaign O

sent an average of 300 million e-mails aday with an average read rate of 15 percent. These fundraising campaigns wererun concurrently and the success ofCampaign O dwarfed that of CampaignR, raising a staggering US$600 million.3

Simply by increasing broadcast reachthrough frequency at near zero marginalcost, Campaign O, or Campaign Obamato give it its full name, trounced CampaignR (for Romney) and produced the mostsuccessful digital fundraising campaign ofall time.

References1. Alchemy Worx (2013) ‘How the Obama campaign

succeeded with low open rates’, available at:http://www.alchemyworx.com/emailworx/2013/strategy/how-the-obama-campaign-succeeded-with-low-open-rates (accessed 18th March, 2013).

2. Connelly, L. (2011) ‘View from the digital inbox2011’, available at: http://www.merkleinc.com/thought-leadership/white-papers/2011/view-digital-inbox-2011 (accessed 18th March, 2013).

3. Bloomberg BusinessWeek (2013) ‘The sciencebehind those Obama campaign e-mails’, availableat: http://www.businessweek.com/articles/2012-11-29/the-science-behind-those-obama-campaign-e-mails (accessed 7th February 2013).


Quist


practice papers beyond open rate: why it is time for e ... · alchemy worx, 4th floor, 1 new oxford...

Documents