Contribution of the Alive & Thrive-UNICEF advocacy efforts to improve infant and young child feeding policies in Southeast Asia.

Abstract Evaluating the impact of advocacy for policy change presents many challenges. Recent advances in the field of evaluation, such as contribution analysis (CA), offer guidance on how to make credible claims regarding such impact. The purposes of this article are (a) to detail the application of CA to assess the contribution of an advocacy initiative to improve infant and young child feeding policies and (b) to present the emergent theory of change and contribution story of how progress was achieved. An evaluation applying developmental evaluation and CA was conducted on the Alive & Thrive (A&T)–UNICEF initiative in seven Southeast Asian countries to document the extent to which policy objectives were achieved and identify key drivers of policy change. A contribution story was developed based on these experiences. The advocacy approach, which involved a four‐part process, contributed directly to (a) set the agenda of various actors and (b) create a strategic group; and indirectly to (a) set and maintain the issue on the agenda at all stages of the policy cycle, (b) support the government to carry out a set of critical tasks, and (c) extend commitment. All of this helped to achieve progress towards policy change. External influences were at play. The flexibility of A&T allowed key actors to utilize the positive external influences and address some of the negative ones through developing responsive strategies mitigating their effects. The emerging contribution story supports that A&T–UNICEF initiative contributed to the progress achieved in the participating countries.

in 1991 by WHO and UNICEF, to foster health systems more supportive of breastfeeding, and recently updated (World Health Organization, 2018). Although those international policy frameworks seek to help countries in creating their own legislation, their existence is insufficient to trigger policy change. Advocacy strategies are needed in order to motivate and guide efforts of various institutions and ensure that countries translate international frameworks into national measures.
Advocacy is often considered more as an art than a science, and advocates require strategic capacity to be able to influence different stakeholders located at critical points within the policy process (Gillespie, Haddad, Mannar, Menon, & Nisbett, 2013;Pelletier et al., 2013;Pelletier, Menon, Ngo, Frongillo, & Frongillo, 2011). Although there is no commonly agreed definition, advocacy is used to achieve social or policy change and implies framing the issue, developing alliances, gathering and disseminating data (Reisman, Gienapp, & Stachowiak, 2007). A fast growing body of grey literature has developed to provide useful tools and guides to assess advocacy efforts (Guthrie, Louie, David, & Foster, 2005;Mansfield, 2010;Reisman et al., 2007;Tsui, Hearn, & Young, 2014). Evaluating the progress achieved thanks to policy advocacy initiatives presents numerous challenges due to: the multiple external factors that influence systems; the time frame (often requires fast adaptation to context but takes long time to produce outcomes); the myriad of actors, audiences, and interactions involved; and the fact that many tactics happen behind closed doors (Glass, 2017). Policy change is also a highly context-sensitive phenomenon, which brings its own challenges (Riley et al., 2017). The following quote illustrates well the set of challenges faced when evaluating policy advocacy: "Evaluators must acquire and accurately weigh and synthesize imperfect information, from biaised sources with incomplete knowledge, under rapidly changing circumstances where causal links are almost impossible to establish" (Teles & Schmitt, 2011;Gardner & Brindis, 2017, p. 76). The quote also points to the problem of attribution (cause-effect), which is well recognized and highly discussed in the sources referred throughout this section.
Recent conceptual and methodological advances in the field of evaluation provide approaches that can help make credible claims about whether and how some advocacy activities contributed to policy changes. For example, contribution analysis (CA) has proved to be very useful in exploring attribution questions when conventional experimental designs cannot be carried out to assess the contribution of an intervention (Mayne, 2001(Mayne, , 2011(Mayne, , 2012. This theory-based plausibility analysis is increasingly considered helpful in assessing the contributions of complex interventions (Dybdal & Lemire, 2010;Patton, McKegg, & Wehipeihana, 2015). CA also has been proposed for the evaluation of advocacy initiative for policy change (Gardner & Brindis, 2017). An increasing but still limited body of literature presents examples of the application of CA to various programmes and initiatives (Biggs, Farrell, Lawrence, & Johnson, 2014;Delahais & Toulemonde, 2012;Kotvojs & Shrimpton, 2007;Mayne, 2012). Several challenges in applying CA have led to a continued refinement of the approach. Although CA appears to now enter into a fourth generation (Budhwani & McDavid, 2017), some challenges still remain in its application. The present article seeks to add to the scarce literature on the application of CA to evaluate policy advocacy initiative, by providing insights from an evaluation that took place to assess the contribution of a complex advocacy initiative.
Alive & Thrive (A&T) implemented a 9-year initiative to improve IYCF policies and practices. A first phase of implementation in Bangladesh, Ethiopia, and Vietnam led to considerable gains in building sustainable and enabling environments for the implementation and scaling up of IYCF programmes . A second phase began in 2014 in which advocacy efforts expanded to include seven Southeast Asian (SEA) countries (Cambodia, Indonesia, Lao People's Democratic Republic, Myanmar, Thailand, Vietnam, and Timor-Leste) and two African countries (Burkina Faso and Ethiopia). The advocacy efforts focused on sharing Vietnam's policy advocacy experiences and supporting organizations in other countries to advocate for either the adoption of IYCF-friendly policies or the implementation, enforcement, or monitoring of existing policies. The work focused on three main policy areas: the Code, maternity protection, and health system strengthening.
The advocacy efforts were based on the main strategies identified in the first phase of the initiative (Alive and Thrive, 2013) that evolved into a four-part process for policy change: (a) establish and sustain partnership, (b) develop evidence base, (c) develop messages and materials, and (d) build consensus. The A&T initiative collaborated closely with UNICEF to undertake the advocacy efforts in the seven SEA countries regarding those policy areas. A real-time evaluation of the advocacy efforts of actors in those countries took place in 2015-2017; the application of CA to this advocacy initiative is presented in this article.
The overall objective of the evaluation was to document the extent to which policy objectives were (or were not) achieved in each country and to identify the key drivers of policy change. More specifically, this evaluation sought to better understand whether and how the activities carried out within the advocacy initiative as part of a regional effort in the seven SEA countries, resulted in policy environments that were more supportive of IYCF. Although this evaluation investigated all three policy areas, the present article focuses solely on the Code. The purposes of this article are (a) to detail the application of CA to assess the contribution of a multicountry advocacy initiative and (b) to present the emergent theory of change (ToC) and contribution story of how progress was achieved.

Key messages
• Global guidance to improve infant and young child feeding policies exist, but advocacy efforts are needed to help countries align with international policy frameworks.
• Evaluating the impact of advocacy initiatives for policy change presents numerous challenges.
• Contribution analysis is a six-step approach that can help explore attribution questions in complex environments.
• Using contribution analysis with developmental evaluation presents several strengths for advocacy and policy change evaluation.
• Applying these two evaluation approaches confirmed that the advocacy efforts of A&T, UNICEF, and partners contributed to progress in Southeast Asia.

| METHODS
This real-time evaluation used a combination of two theory-based evaluation approaches. First, developmental evaluation (DE) was the overall framework for engaging with actors in the field. DE supports the development of an innovation by collecting various types of data to provide feedback to stakeholders and help them adapt the innovation to the emergent and dynamic context (M. Patton, 2010).
Considering that the strategies and actions carried out by A&T, UNICEF, and partners were different in each country and evolved in response to the different contexts, the use of DE was especially relevant. Second, CA allows for the assessment of whether an intervention contributed to the observed effects (Mayne, 2011). CA involves elaborating a postulated ToC of a programme or intervention and testing it, while taking into account other influencing factors.
It offers a systematic way to make evidence-based causal claims in situations that do not lend themselves to conventional experimental or statistical methods. The present article focuses on CA but draws upon the data generated through DE and other sources.

| Developmental evaluation
The DE approach took place between May 2015 and March 2017.
Participants were actors working to improve IYCF in the various countries, representing A&T, UNICEF, government ministries, research institutions, and non-governmental organizations. These participants had been primarily selected by A&T staff because they needed to identify the main actors involved in their advocacy efforts. In addition, the main researcher proposed to consider actors from additional organizations based on the main country priorities. Both of them easily reached agreement on the main actors to interview. Additional information can be found in a companion article of this supplement .
Several data collection methods were used. First, participant observation was done with five A&T staff and representatives (focal point for the various countries). This was possible because the researcher (IML) travelled with the A&T focal points who took advantage of the trip to follow-up on the execution of activities described in the country work plans. The observation of those activities helped gain contextual knowledge and document the various ideas and actions undertaken by different actors. The researcher filled out a template identifying various items for each country: types of activities; objectives; participants; inputs, tactics, or strategies; target audience; progress; and comments or questions. At the end of each trip, the researcher and the focal point took a time for debriefing and discussing some insights. Detailed notes were taken throughout the duration of the trip.
Second, the researcher and the A&T focal point held key informant meetings with a total of 98 actors in the seven countries (Cambodia, n = 11; Indonesia, n = 19; Lao PDR, n = 13; Myanmar, n = 17; Thailand, n = 9; Timor-Leste, n = 13; and Vietnam, n = 16). Considering that the status of the activities was different for each country, the discussions with the actors varied. The objectives of the key informant meetings in country were two-fold: (a) for the researcher to introduce the realtime evaluation and identify several actors who could be interviewed at different times throughout the evaluation period and (b) for A&T focal point to follow-up on the execution of activities and create advocacy strategies with the country actors. A total of 13 meetings were taperecorded during the country visits. Tape-recording was possible when the actors and A&T staff had already developed strong relationship and when both the researcher and the A&T focal point felt that it was appropriate to tape-record. Although not all meetings were taperecorded, the researcher took detailed notes during and after those meetings. After each meeting, the researcher also debriefed with the A&T focal points to ensure having captured the most important points (especially when there was translation involved).
Third, in-depth interviews were held with 28 actors throughout the real-time evaluation (Cambodia, n = 3; Indonesia, n = 5; Lao PDR, n = 4; Myanmar, n = 5; Thailand, n = 3; Timor-Leste, n = 5; Vietnam, n = 3). DE requires engaging with key actors during the development of their innovation. Those interactions served as data collection opportunities to generate insights based on their experiences. The number of actors and interviews per country depended on several factors: intensity of the A&T work in country, actors previously met during country visits, accessibility to some of them (e.g., accessibility was more difficult with government actors who did not speak English), involvement in A&T-UNICEF strategies, availability, and other logistic considerations. A total of 44 interviews were tape-recorded.
Finally, a desk review was conducted with a broad diversity of documents, including research and strategic documents, A&T working documents, reports, and progress updates. A large number of documents specific to the Initiative were made available in a dropbox file by A&T and could be consulted by the A&T focal points. One of the researchers had access to this file and could select the ones to include in the analysis. The documents often varied by country, but when a document was available for most countries, those were analysed in a systematic way (e.g., the legal reviews or opinion leader assessments). Other documents came from the A&T website or were shared by A&T focal points. Most documents available on the dropbox had been read in details by one researcher (IML) and helped develop a chronology of events and strategies for each country.
All tape-recorded interviews were transcribed verbatim. Thematic content analysis (Miles & Huberman, 1994) was carried out through an iterative process in which different types of coding (Saldaña, 2012) were performed with the QSR International's NVivo 11 software.
Initially, two researchers (MG and IML) coded individually with cross-checking (double-coding) and ongoing discussion on the codebook.
Once most categories had been developed, coding proceeded individually (MG) with discussion with the research team whenever needed.

| Contribution analysis
The contribution story of a complex intervention can be developed by carrying out iterative steps, as summarized in Annex S1 in the Online Supplemental Materials (OSM).
Step 1: Clarification of the cause-effect issue to be addressed This step involves recognizing the attribution problem and determining the specific cause-effect question to be addressed. It was acknowledged that the objectives of this real-time evaluation involved evaluating whether the advocacy efforts carried out by A&T, UNICEF, and partners contributed to IYCF policy changes. At the onset of the evaluation, underlying questions were formulated to refine the cause-effect questions, as shown in Table 1. This shed light on the Initiative-related complexity and its implications for the evaluation.
Although the underlying evaluation questions were applied to the three policy areas, the CA was carried out only on one policy area (the Code) because the actors in-country focused primarily their work on this issue.

Steps 2 and 3: Develop the ToC and gather existing evidence
In a typical CA, a ToC is developed prior to data collection and confronted to available information regarding an intervention. In the present evaluation, a logic model was initially developed based on the experience of Vietnam and shared with A&T stakeholders from Headquarter and the regional office located in Vietnam. A ToC was later developed based on this logic model and included a list of assumptions.
Step 4: Assemble and assess the contribution story At this stage, a first contribution story can be drafted and critically assessed based on the available evidence. This step usually guides the subsequent collection of data towards the strengthening of the weakest parts of the contribution story. In the present case, data were collected through DE during almost the entire advocacy initiative.
Then, the contribution story was developed around the end of the evaluation with data covering almost all aspects of the postulated ToC.
Step 5: Seek out additional evidence The contribution story drew upon data generated through DE.
Two coding schemes were applied and helped shaping the contribution story, as described below.
Systematic tracking of activities and accomplishments, and documentation of contexts: A systematic tracking of activities and a review of all available documents (desk review, trip reports, country progressions in the interim report, etc.) allowed us to develop emergent and detailed chronologies for each country, which documented the following elements: actors, strategies, events, challenges, contextual factors, and accomplishments. Regarding accomplishments, one challenge was that they were not uniform across the countries. Study of the implementation of the A&T advocacy approach: The data collected throughout the evaluation were coded and specifically analysed in regard to the four parts of the A&T advocacy approach (Annex S2 in OSM) to understand how actors carried out each part.
The accumulation of data collected from all countries brought a robust overview of the diverse strategies used by the actors.
Step 6: Revise and strengthen the contribution story At this step, the new data gathered usually help to build a more credible contribution story. Once done, the analysis may return to Step 4 to strengthen the contribution story. After developing our contribution story, this same iterative process helped us refine the contribution claims towards the end of the evaluation.

| Validation of emergent findings
Over the course of the evaluation, two mechanisms helped validated the emergent findings related to the advocacy approach. First, several documents were created and shared with key actors to gather their perspectives. As an example, a midterm report was produced with preliminary insights for each country on the following: context, policy objectives, and activities; strategies and accomplishments; and Main evaluation question Did the advocacy efforts carried out by A&T, UNICEF, and partners in each country contribute to IYCF policy changes?
Underlying questions • What advocacy activities, strategies, and tactics took place in each country?
• How were they carried out?
• What conditions and factors influenced the process of policy change?
• What were the policy changes?
• Were the policy objectives reached?
• How was progress achieved?
• What were the key drivers of the policy changes?
Initiative-related complexity • The national efforts involved a range of emergent (not predetermined) activities, with actors often responding to opportunities and threats.
• The Initiative was carried out in seven countries representing diverse contexts.
• The outcomes were assessed as progress within a policy cycle, rather than an exclusive focus on official policy changes.
• Considering the long timespan required for progress within a policy cycle, intermediate outcomes were captured and acknowledged.
Implications for the evaluation • Activities needed to be tracked prospectively.
• An understanding of the broad and specific policy process in each country was required.
• The multiple contexts altogether needed to be accounted for.
• Progress depended on the country's stage at the beginning of the Initiative and on the stage reached.
• The results chain involved a large number of linkages between all the elements.
1 The stages enumerated here come from a paper of this supplement that specifically relates to the Code (Michaud-Létourneau et al., 2019). Although we use the term "stage," we acknowledge that these processes are iterative, overlapping, and nonsequential. Nonetheless, this distinction highly helps to locate findings and facilitate understanding of the policy cycle. A deeper discussion goes beyond the scope of this paper, but one can further read on the decision process in Auer (2017).
selected insights on the policy process, challenges, and next steps. The content of each of the country briefs had been validated by some of the actors interviewed in country. As another example, towards the end of the study, when emerging findings on the Code had been articulated, the researchers presented them to some of the most engaged policy advocates (from A&T, UNICEF, and civil society) to gather their reaction to the emerging findings. Changes were then made to integrate their comments. Second, before each interview with actors in country, documents and data available for the country were reviewed, and questions for cross-checking with interviewees were identified, in addition to validating emergent insights.

| Initial ToC
In Step 2 of the CA, an initial ToC was developed from a logic model of the intervention based on the experience in Vietnam. It involved the four-part advocacy approach used by A&T, UNICEF, and partners ( Figure 1). Three linkages are illustrated from the outputs to the impact, along with a set of assumptions for each of those. Although the same advocacy approach was applied in each of the seven countries, the exact activities were not predetermined at the onset of the evaluation because, as it is usually the case for advocacy, actors respond to opportunities in each country.

| Final contribution story
The development of the final contribution story required verifying the various assumptions related to the three main linkages identified in the ToC. The pathway by which progress was achieved is described below, with some quotes to illustrate the contributions.

| Linkage 1: Activities/outputs to proximal outcomes
The first linkage connects the activities/outputs to proximal outcomes.
The assumptions for this linkage were investigated in regard to each part of the advocacy approach, and the evidence gathered is presented for each part. The first part of the advocacy approach was to build evidence. A&T initiated or supported numerous studies and analyses to develop arguments and convince actors from the government on the importance of taking action. Doing the right studies to get the right evidence at the right time was critical to advance the work, as expressed by this actor: … when A&T came in, they looked at what they could do. Researcher in an academic institution, 2016.06.10 Actors from government and development organizations interviewed during this real-time evaluation applauded the way by which those two organizations worked together and their complementary strengths.

A strategic group of actors
Another key aspect of the partnerships and major contribution of the advocacy approach to progress was through the creation of a strategic group of actors. In a given country, various groups were already at play. To be honest, anyone who looks at The Lancet or who has exposure to this kind of discussion would never be in a position to say that "this is something that's impossible or challenging, or that this is something that does not make sense." This is something that's needed.
[…] The idea of having to implement these laws is to ensure that you get the best out of the nutrition investment, which is a long-term economic investment over the little amount of time. Because here's what was presented: Every $1 dollar or every $1,000 dollars the industry spends to improve the marketability of their products, government cannot compete, not even below $1 dollar. It might be better for the government to take a stance of, "We do not want to compete. We just want to curtail and limit what you (the industry) can do." And that's the general message.
Government representative, Indonesia, 2016.11.06 The quote above is from an individual who was invited to the 2016 A&T-UNICEF regional policy workshop and who then became a champion for the cause and became a catalyst for change in his country. Large events were opportunities to raise awareness and influence many key actors in high-level, well-organized, and visible venues.
In conclusion, each part of the advocacy approach has contributed significantly to the specific outcomes presented above. Nonetheless, taken as a whole, one main contribution of the advocacy approach has been to put the Code on the agenda of various key stakeholders.
This section has presented evidence on various contribution claims confirming the first set of assumptions. It has shown that key policy makers from the government were reached. They understood the importance of the Code and the role they could play, and some of them engaged themselves in moving the Code forward. This first linkage is depicted in detail in Annex S3 (OSM).

| Linkage 2: Proximal outcomes to distal outcomes
The second linkage connects the proximal outcomes to the distal outcomes. The assumptions examined related to capacity change and policy progress, which involved increasing the capacity of policy makers to develop effective IYCF policies and work towards putting them in place.

A set of critical tasks
The distal outcomes were conceptualized as progress within the policy cycle, that is, an advancement from one stage of the policy cycle to the following one. Those stages involved development of the Code; adoption; preparation for implementation; monitoring and enforcement; and evaluation, learning, and adaptation. For each of these stages, activities were identified. Those were carried out by the seven governments, supported by the strategic groups and linked to progress within the policy cycle. These activities were later termed "critical tasks": none of them was sufficient to influence the whole policy cycle by itself, but governments carrying out those activities were more likely to make progress within the various stages of the policy cycle. Thus, the distal outcomes, represented as a unique box in Figure 1, involve the realization of the critical tasks. The engagement of key relevant actors to support the governments One major contribution of the strategic group was to bring a network of people and a pool of resources to advance the work. The strategic group was able to engage key actors needed at various stages of the policy cycle. First, the engagement of key local actors helped maintain the Code on various agendas. Second, the engagement of international experts was used to support the governments in carrying out the critical tasks and achieve progress within the policy cycle. Third, the strategic group was able to extend commitment by engaging with actors beyond the health sector and with civil society.
Regarding specifically the policy-making process, A&T and UNICEF played the role of bringing technical expertise on the Code at different times, for example, for drafting the various pieces of FIGURE 2 Groups of actors involved in the work on the Code legislation or for providing solid arguments to counter the lobby from the industry during public hearings. By engaging the right actors to move the work forward within the policy cycle, A&T, often with UNICEF support, were "connecting the dots" within the countries.
A&T was often able to fill gaps that no one could fill in a timely manner, bringing a significant contribution to the overall efforts to move the processes forward, as supported by this actor: Sometimes, the windows of opportunity are very short, it's like opening and closing, and you find it out, and that's it, so it's important to have support and expertise available to be able to take advantage of these windows.
[…] Not only funding but flexibility … because even if funding is there, for example, in organizations like UNICEF … because you need to be able to … Okay, you went to the government, talked, and asked: "it's (for) tomorrow, next week, in 3 weeks … no more than 2 months." You need to be able to respond dynamically.
It depends on where you are in the policy process, but UNICEF, it would take 4 to 6 months, just, institutionally, to say, "Okay, we do that." … Okay, A&T seems to have the flexibility of inserting this.
UN agency representative, 2016.02.04 In summary, as seen in the examination of the first linkage, the advocacy approach led to the creation of a strategic group of actors.
In the second linkage, this strategic group helped set and maintain the Code on various agendas and was able to engage various key actors to facilitate the undertaking of the critical tasks and progress within the policy cycle. This led to policy makers being assisted to design effective IYCF policies and submit them for adoption. The governments were also supported to develop strong mechanisms to implement, enforce, and monitor those policies. This second linkage is illustrated in detail in Annex S3 (OSM).

| Linkage 3: Distal outcomes to impact
The third linkage connects the distal outcomes to impact.

| External influences
The creation of a credible contribution story requires attention to the potential influence of contextual factors. According to the latest thinking on contextual factors, Mayne who articulated CA, uses the term external influences and define it as the "social and economic trends at work, specific events that have occurred, other interventions overlapping with the intervention in question (with similar aims) or environmental factors at work" (Mayne, 2018, p. 5). In A&T-UNICEF advocacy initiative, several external influences appeared to have hindered or facilitated the advocacy efforts. These influences affected the work at various times, and it was possible to link them to the different set of assumptions. Figure 3 presents an aggregate list of these external influences across the seven countries and presents them according to the linkages of the ToC.
As mentioned earlier, an important feature of advocacy-related initiatives is that many of the strategies employed by the actors are not predetermined and emerge depending on the context (Gardner & Brindis, 2017). The A&T-UNICEF initiative is no exception. The strategic group of actors had a strong capacity to analyse the contexts and were able to develop their strategies accordingly, guided by the advocacy approach. Thus, when contextual factors affected the advocacy efforts, the actors were also able to adapt to those external influences. They took advantage of the positive external influences and tried to mitigate the negative ones. Table 3

| DISCUSSION
The use of CA helped demonstrate the contribution of the advocacy efforts carried out by A&T, UNICEF, and partners to progress regarding IYCF policies in the participating countries and revealed how progress was achieved. As proposed by Befani and Mayne (2014): "if one can verify or confirm a ToC with empirical evidence-that is, verify that the steps and assumptions in the intervention ToC were realised in practice, and account for other major influencing factors-then it is reasonable to conclude that the intervention in question has made a difference, i.e. was a contributory cause for the outcome." Applying the main steps of CA described by Mayne allowed us to verify the assumptions of the postulated ToC and to take into account the external influences. Thus, a plausible contribution story was developed and supports that A&T-UNICEF advocacy efforts with partners made a difference and contributed to the progress achieved in the various countries. The CA also revealed how progress was achieved. The four parts of the advocacy approach helped set the agenda of various actors and allowed the creation of a strategic group, which represented a main driver of progress.
Therefore, this advocacy approach was very effective, and public • Approval of TOR for the oversight board, the executive working group, and the control committee.
• Approval of Guidelines for the implementation, monitoring and enforcement of Sub-Decree 133 and Joint Prakas 061.
April-May 2016 • The government and partners developed four checklists (one per line ministry). A 1-week workshop was carried out to get consensus on the content.

August 2014
• Creation of an oversight board that led to the creation of a proper mechanism to ensure enforcement. A focal point from each line ministry was assigned to it.
• Creation of the control committee and the executive working group.

May 2016
• Consensus on the checklists for the monitoring of the Code • Preparation for the monitoring of the Code Indonesia March 2016 • The launch of the Lancet Series on Breastfeeding and the presentation of the "Cost of Not Breastfeeding" took place during the National Nutrition Day-identification of high-level champions.

May 2016
• MOH attended the meeting for WHA Resolution 69.9 and the Government of Indonesia has endorsed this resolution.

January 2017
• Five health and nutrition organizations launched a joint statement regarding WHA Resolution 69.9.
August 2016 • World Breastfeeding Week during which the Director General mentioned in front of a large forum that the Government of Indonesia supported the WHA Resolution 69.9.
December 2016 • The Food Standardization Unit (under the BPOM) was working on the revised draft of the Code.

Lao PDR November 2014
• Inter-Parliamentary Union event held in Vientiane: One of the priority actions was the BMS Code.

July-October 2016
• MOH has organized two meetings with key stakeholders (UNICEF, SC, and other line ministries to talk about the Code).
• A draft of TOR for a Task Force for the Code is developed.

January 2017
• First meeting of the Task Force took place in which a first draft of the revised Code is discussed.

November 2014
• High-level commitment to strengthen the Code January 2017 • Draft of a Prime Minister's decree for the Code

Myanmar December 2015
• The government disseminated the Order (their Code) to the formula companies in a workshop.

March 2016
• SC and SUN-CSA monitor a database of BMS Code violations (through KoboCollect) and submit routine reports to government.

May 2016
• Second official meeting of the TWG took place. A deadline for "voluntary recall" of products violating the BMS Code was set.
• Monitoring reports begin to be produced and sent to the TWG.

2014
• A National Order of Marketing of Formulated Food for Infant and Young Child was approved under the National Food Law.

November 2015
• A National TWG was established as the official national government body charged with the design and oversight of the overall process for monitoring and enforcing the National Order.

2016
• Official deadline for "voluntary recall" of products violating the Order (Code): 24 July 2016.
• Revision of the deadline to November 2016.
Thailand August 2015 • The revised draft circulated between different ministries for comments before going to the Cabinet.

January 2016
• The Global Nutrition Report was launched in Bangkok, showing that Thailand was off course on all the WHA nutrition indicators. Surprised by those, the MOH committed to take action.

December 2015
• The State Council has approved the draft of the BMS Code as a law (with the ban for up to 24 months).

2016
• Commitment of the MOH to improve the breastfeeding rates 2017 • BMS Code Act passed with more than 90% vote

Timor-Leste
April 2016 • Regional Policy Workshop in Bangkok: The BMS Code was raised as a priority among the country team members who attended.

None captured
Vietnam December 2014 • Three national dissemination workshops took place to train and inform everyone on the content of the Decree 100.

September 2016
• The Ministry of Planning and Investment proposed a revision to Article 7.4 in the 2012 Advertisement Law to narrow the ban on advertisement of BMS for children from 24 to 12 months of age. The Ministry of Planning and Investment proposed using the fast-

December 2014
• Decree 100/2014/ND-CP on marketing and use of feeding products for young children, feeding bottles, teats, and pacifiers was approved. The decree further specifies the Advertisement Law.

2015
• Trainings were carried out countrywide for the monitoring of the Code.
(Continues) advocates trying to inform IYCF policies can find more information for using it (Alive and Thrive, 2016). The contribution story also uncovered the support brought by the strategic group for the realization of a set of critical tasks, previously identified as a main driver of progress within the policy cycle (see ). The use of the advocacy approach by the strategic group of actors also ensured setting the agenda for the Code and maintaining it there at all stages of the policy cycle. All of this was crucial to achieve progress towards policy change.
Applying CA to this initiative also brought several insights:    Of relevance, it is important to consider that becoming acquainted with DE and CA requires practice, time, and engagement, and allows for trials, errors, and adaptations. Funders also need to understand the need to engage with advocates and people in the field if they are to support this type of evaluation.
Regarding limitations, the fact that external influences were identified throughout the whole data collection but combined mostly at the end of the evaluation can be perceived as a limitation.
However, this helped avoid biases from focusing only on a limited number of selected factors. Another potential limitation of this study may be the limited evidence gathered around the third linkage.
The literature recognizes that the more we move along the impact pathway, the less evidence we are able to gather. This refers to the sphere of influence of an intervention. Indeed, moving from outputs to impact involves going from direct control (operational environment) to direct influence and eventually indirect influence (Montague, Young, & Montague, 2003).
Overall, the credibility of the contribution story is strengthened by having investigated seven countries. This is because, regardless of contexts, the detailed ToC (created based on the data collected in all countries) is seen to apply well to each country individually. The countries where the advocacy efforts were less developed also provided counter factual to reinforce the importance of some elements breastfeeding. Therefore, the application of the findings of this paper and its supplement goes beyond the Code.

CONTRIBUTIONS
IML designed the real-time evaluation, played a leadership role throughout all stages of the study, collected most data, and conceptualized the paper.
IML and MG conducted data analysis, interpretation of results, and drafting of the different sections of the manuscript. DLP participated in data collection, advised at all stages of the process, and collaborated in revising earlier drafts of the manuscript. All authors read, commented, and approved the final manuscript.